EP2227750A1 - Systems and methods for intelligent paperless document management - Google Patents
Systems and methods for intelligent paperless document managementInfo
- Publication number
- EP2227750A1 EP2227750A1 EP07865320A EP07865320A EP2227750A1 EP 2227750 A1 EP2227750 A1 EP 2227750A1 EP 07865320 A EP07865320 A EP 07865320A EP 07865320 A EP07865320 A EP 07865320A EP 2227750 A1 EP2227750 A1 EP 2227750A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- documents
- document
- instant invention
- user
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/93—Document management systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/26—Techniques for post-processing, e.g. correcting the recognition result
- G06V30/262—Techniques for post-processing, e.g. correcting the recognition result using context analysis, e.g. lexical, syntactic or semantic context
- G06V30/268—Lexical context
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Definitions
- Tlic field of invention is the process or method of doing business by processing paper documents, image files, and/or electronic documents; using a computer to store, retrieve, display, analyze, collate and capture information from the documents; optionally using a computer for making decisions based on this information; and exchanging the organized information between organizations electronically.
- the field of the invention also includes the method of performing such analysis, collation, and information capture as well as an apparatus for conducting such analysis, collation, and capture.
- such systems are limited to document boundary detection, document classification and text extraction and do not offer advanced document collation with separation of very similar documents, and domain-sensitive scrubbing of extracted information into usable data;
- the documents related to a given loan for the purchase of a property or properties in any jurisdiction may be assembled into electronic images by scanning (or direct entry, if already in an electronic form) before, during and after funding of the loan to form a partially, or preferably, complete document set, referred to herein as the "Dox Package.”
- These documents originate from a number of sources, including banks and/or credit unions.
- the order of these documents are assembled and filed depends very much on the individuals involved, their timeliness and their preferences, organization, or disorganization in sorting the various forms and other documents containing the required information.
- the systems and methods described have at least one of the following features: (a) receiving paper in any format typically used to transfer paper digitally between one human/system to another without requiring the sender to prepare or separate the Dox Packages with coded cover sheets or separator sheets; (b) using at least one auto-indexing as described herein to convert that raw feed into collated, named, indexed documents minimizing human interaction; (c) identifying and making available data points from those documents; (d) vaulting the digital paper and data in a commercially secure fashion; (e) filing documents using contextually based references so that they can be located by those who need to work with them; (f) making available rich, usable interfaces displaying a digital, paperless workplace accessible by different devices and types of users across vast geographies; (g) reproducing paper-based workflows and enhance.
- the systems and methods described have at least two of the features listed above. In some embodiments, the systems and methods described have at least three of the features listed above. In some embodiments, the systems and methods described have at least four of the features listed above. In some embodiments, the systems and methods described have at least five of the features listed above. In some embodiments, the systems and methods described have at least six of the features listed above. In some embodiments, the systems and methods described have at least seven of the features listed above.
- the systems and methods described provide a Web-based solution which enables intelligent document paperless management and collaboration, eliminating inefficiency caused by paper documents.
- the systems and methods described herein provide Web-based paperless offices where users can collect, store, and share all documents from various locations.
- the systems and methods described herein can receive images of documents from any source, such as e-mail, fax, ftp upload or scanner/digital copier, then place them into electronic folders, where they can be viewed and acted upon exactly as they were in paper form.
- the imaged documents reside in a warehouse repository and an administrator allows permitted viewers' access to the documents based on their individual role in the process, using links via the Web.
- the documents do not travel at all, but everyone needing access to them can have the documents at their fingertips.
- the systems and methods described herein can provide hierarchical virtual storage such as virtual offices and virtual cabinets that can be flexibly setup to best represent an existing organization's file storage structure.
- These virtual (electronic) cabinets can be set up according to various factors to best fit for the user.
- a mortgage broker can setup cabinets/offices by loan officer, loan type, workflow, or time of origination. Documents can be automatically indexed and placed into electronic folders within these electronic cabinets.
- the systems and methods described herein have at least one of the following functions: (a) can perform automatic indexing and document classification as images arc received, (b) can provide imaged documents residing within secure and easily accessed environment, and (c) can allow full collaboration among permitted stakeholders all along the value chain.
- the systems and methods described are Web-based and have all three functions listed above.
- Web-based systems for intelligent paperless document management comprising,
- said at least one automatic indexing engine does not require barcodes or separators.
- said system does not provide the capability of prearranging access levels according to functions of said multiple users.
- said system docs not require the capability of prearranging access levels according to functions of said multiple users.
- said system does not prohibit a user from modifying or deleting a document in said depository.
- said at least one automatic indexing engine excludes involvement of barcodes or separators for the purpose of indexing or classifying documents.
- barcodes can be used for folder IDs.
- said system can forward incoming documents into folders designated by the folder IDs.
- said system excludes involvement of prearranging access levels according to functions of said multiple users.
- said system provides a user capability to modify or delete a document in said depository.
- system described further comprises at least one electronic cabinet. In some embodiments, the system described further comprises at least one electronic folder. In some embodiments, the system described further comprises a delivery center to allow custom delivery packages comprising different documents for multiple parties.
- the system described further comprises at least one business rule manager. In some embodiments of the systems comprising at least one business rule manager, said business rule manager uses at least one workflow rule. (0026J) In some embodiments, the system described further comprises at least one data extraction engine. In some embodiments of the systems comprising at least one data extraction engine, the systems have capability to output extracted data in a specified format. In some embodiments of such systems, the specified format is MISMO SMART Doc XML.
- systems comprising at least one data extraction engine, wherein the systems have capability to flag for human intervention when inconsistent data is identified; said human intervention include sorting, modifying, and/or deleting documents.
- the online collaborative portal allows simultaneous access from multiple users.
- the system comprises at least two user accounts corresponding to each of at least two inboxes.
- the inbox accepts incoming documents selected from the group consisting of e-mails, faxes, ftp uploads, scanners, digital copiers, and combinations thereof.
- the system described further comprises a desktop tool for a user with a user account and inbox providing for uploading of documents to said user's inbox.
- the system described can integrated with at least one desk-based system such as mortgage loan origination software (LOS), automatic underwriting systems (AUS), lender underwriting systems (LUS), automate mortgage compliance software, or other electronic loan software.
- LOS mortgage loan origination software
- AUS automatic underwriting systems
- LERS lender underwriting systems
- automate mortgage compliance software or other electronic loan software.
- step (c) indexing documents received in step (b) with at least one automatic indexing engine
- said at least one automatic indexing engine does not require barcodes or separators. In some embodiments of the above aspect, said method does not provide the capability of prearranging access levels according to functions of said multiple users. In some embodiments of the above aspect, said method does not require the capability of prearranging access levels according to functions of said multiple users. In some embodiments of the above aspect, said method does not prohibit a user from modifying or deleting a document in said depository.
- said at least one automatic indexing engine excludes involvement of barcodes or separators for the purpose of indexing or classifying documents.
- barcodes can be used for folder IDs.
- said method can forward incoming documents into folders designated by the folder IDs.
- said method excludes involvement of prearranging access levels according to functions of said multiple users.
- said method provides a user capability to modify or delete a document in said depository.
- the method described further comprises the step of placing indexed documents into at least one electronic cabinet. In some embodiments, the method described further comprises the step of placing folders into at least one electronic cabinet. In some embodiments, the method described further comprises the step of placing indexed documents into at least one electronic folder. In some embodiments, the method described further comprises the step of delivering documents to at least one person using a delivery center to allow custom delivery packages comprising different documents for multiple parties.
- the method described further comprises the step of organizing documents using at least one business rule manager.
- said business rule manager uses at least one workflow rule.
- the method described further comprises the step of extracting data using at least one data extraction engine.
- the method described further comprises the step of outputting extracted data in a specified format.
- the specified format is MTSMO SMART Doc XML.
- the method described further comprises the step of flagging for human intervention when inconsistent data is identified; said human intervention include sorting, modifying, and/or deleting documents.
- the online collaborative portal allows simultaneous access from multiple users.
- the inbox accepts incoming documents selected from the group consisting of e-mails, faxes, ftp uploads, scanners, digital copiers, and combinations thereof.
- the method described further comprises the step of providing for a desktop tool for uploading documents.
- the online collaborative portal can integrated with at least one desk-based system such as mortgage loan origination software (LOS), automatic underwriting systems (A(JS), lender underwriting systems (LUS), automate mortgage compliance software, or other electronic loan software.
- Figure 2 depicts a detailed diagrammatic view of the business methods of the instant invention, i.e., the system flow of a preferred embodiment of the invention.
- Figure 3 depicts an embodiment of the Document Learner process, i.e., the flow of the classification learner.
- Figure 4 depicts an embodiment of the Business Object formation elements.
- Figure 4A depicts an embodiment of the relationship of Knowledge Objects within a Business Object.
- Figure 4B depicts the process of Dox Package creation in one embodiment of the invention.
- Figure 4C depicts the process of document creation in one embodiment of the invention.
- Figure 4D depicts MISMO transaction data-set creation in one embodiment of the invention.
- Figures 5 A and 5B depict screen shots of output obtained through the use of the instant invention. That is, using a Dox Package analyzed by the method/apparatus as described herein, the exemplary data in the figure was available for analysis.
- Figure 6 shows an exemplary embodiment of the network deployment of intelligent paperless document management (IPDM).
- Figure 7 shows an exemplary embodiment of the administration of intelligent paperless document management.
- Figure 8 shows an exemplary embodiment of the client deployment of intelligent paperless document management.
- IPDM intelligent paperless document management
- Figure 9 shows the sign-up site for setting up a new Client Site for a new company using katalyst.net.
- Figure 10 shows that each company can have a unique name. If a particular name has been used, the program will ask the user to input a different name.
- Figure 1 1 shows that a new company name has been accepted by the program.
- Figure 12 shows that the user must have a unique site address in order to login as the first step.
- Figure 13 shows that the user can input info for the company as the second step.
- Figure 14 shows that the user can input billing info as the third step.
- Figure 15 shows that an administrator can be designated as the fourth step.
- Figure 16 shows that the user can choose various subscription details as the fifth step.
- Figure 17 shows that the user confirms the end user agreement as the sixth step.
- Figure 18 shows that the user clicks to yes button to confirm the end user agreement.
- Figure 19 shows the message stating that the set-up process has been completed.
- Figure 20 shows that the desktop tool Messenger has three major functions - scan, print, and pdf upload files for uploading documents into the Web-based system.
- Figure 21 shows the scan function of Messenger.
- Figure 22 shows the print function of Messenger.
- Figure 23 shows the pdf upload function of the Messenger.
- 100661 Figure 24 shows that a folder can be created or designated for uploading documents via scan, print, or pdf upload.
- Figure 25 shows that a demo document is being uploaded and bypassing auto-indexing.
- Figure 26 shows that the user can monitor the progress of uploading the document.
- Figure 27 shows that the demo document has been successfully uploaded.
- Figure 28 shows that the company info can be viewed after logged into the loan katalyst.
- Figure 29 shows the delivery center of loan katalyst.
- Figure 30 shows that delivery can be carried out via fax or e-mail (see Figure 29).
- Figure 31 shows a user can configure a delivery using loan katalyst.
- Figure 32 shows that delivery can be carried out via flp.
- Figure 33 shows that a user can create a new cabinet using loan katalyst.
- Figure 34 shows that a user can create a new document type using loan katalyst.
- Figure 35 shows the inbox information and inbox messenger of loan katalyst.
- Figure 36 shows that a user can create a new office using loan katalyst.
- Figure 37 shows that a user can create a new role using loan katalyst.
- Figure 38 shows that a new user can input details about himself or herself using loan katalyst.
- Figure 40 shows that a user has access to various offices and cabinets using loan katalyst.
- Figure 41 shows that the program is uploading documents and the process can be monitored.
- Figure 43 shows that picture files can be easily uploaded and viewed using loan katalyst.
- Figure 44 shows that loan katalyst allows thumbnail view for documents.
- Figure 45 shows that a user can create a custom delivery package using loan katalyst.
- Figure 46 shows that the user selects documents for the custom delivery package using loan katalysl.
- Figure 47 shows the details of the custom delivery package before delivery using loan katalyst.
- Figure 48 shows that multiple recipients can be chosen using loan katalyst.
- Figure 49 shows that the order of each document can be adjusted before delivery using loan katalyst.
- Figure 50 shows loan katalyst allows the use of password for the custom delivery package for security reasons.
- Figure 51 shows the message stating the delivery was successful using loan katalyst.
- Figure 52 shows an example of fax cover sheet including detail information for a designated electronic folder.
- Figure 53 shows a user can view the document image and extracted data at the same time using loan katalyst.
- Figure 54 shows the incoming transmission log of the subject folder using loan katalyst.
- Figure 55 shows the sent transmission log of the subject folder using loan katalyst.
- Figure 56 shows that a user can add a new note for a document using loan katalyst.
- Figure 57 shows that a user such as a broker can upload documents into a particular electronic folder using loan katalyst.
- Figure 58 shows that the user can choose to bypass auto-indexing and can monitor the process of uploading a document using loan katalyst.
- Figure 59 shows the "indexing option" button for explaining auto-indexing using loan katalyst.
- Figure 60 shows the explanation of bypassing auto-indexing using loan katalyst.
- Figure 61 shows that the document is completely uploaded into the system using loan katalyst.
- Figure 62 shows that a particular user can access documents in two folders here using loan katalyst.
- Figure 63 shows that a particular user can access various offices and cabinets using load katalyst.
- Figure 64 shows the website screen for inputting access code to log in the system described.
- Figure 65 shows that the system described provide user guide and FAQ for online help as well as contact info for a customer center.
- Figure 66 shows that a desktop tool can Messenger can be downloaded from the Web-based system,
- a Dox Package may consist of at least two pages, at least three pages, at least five pages, at least ten pages, at least twenty pages, at least fifty pages, at least one hundred pages, or more. Further, as used herein, a Dox Package includes sets of documents in which all the information/data contained therein is not readily available in electronic or digital forms. Thus, a Dox Package may consist of a variety of documents some of which are electronic documents but some of which are paper copies only, or images, such as PDFs or TIFFs 3 of such paper documents.
- the instant invention in some embodiments, can extract the information from the heterogeneous set of documents that forms Dox Package and enter that information into a computer database much faster than, and in some embodiments, with minimal or no intervention from, a human operator; in some cases ten times as fast, twenty times as fast, thirty times as fast, forty times as fast, fifty times as fast, or more. Additionally, in one embodiment, the instant invention can extract and enter information from a Dox Package with human review of, at most, one page in ten, one page in twenty, one page in thirty, one page in forty, one page in fifty, one page in sixty, one page in seventy, one page in eighty, one page in ninety, one page in one hundred, or one page in over a hundred.
- a "Knowledge Object” is a matrix of the information and its association with reference to a particular business process. When a Knowledge Object is not specific to a process and/or a complete domain, it can be cluster of information. Knowledge Objects are intended to be useful and available for decision-making.
- the term "Knowledge Object,” as used herein, refers to a set of facts preferably along with their relationship and association with other Knowledge Objects in a given Dox Package. Knowledge Object is a matrix of relevant information entities such as facts, image field coordinates, value type, intended to address and assist decision making in businesses.
- a "Business Object” is a collected and organized set of information extracted from a Dox Package intended for a business purpose and ready to use to illustrate relationships and/or the utility of Knowledge Objects. It gives a business-centered view of the extracted and organized knowledge for the decision-making process.
- An example of a Business Object is a MISMO standard SMART Doc
- the term "Dox Package” refers to the pile, stack, or file of documents that is delivered, handed, and/or made available to the operator of the instant invention.
- the Dox Package comprises mortgage documents and documents in support of a mortgage, or secondary financing thereof.
- 'Taxonomy refers list of document types (or document classes) expected in any Dox Package. Documents within the Dox Package or taxonomy may consist of multiple pages, but all pages are preferably logically related to the reference page (as defined below).
- the term "escalation” as used herein refers to a subroutine within the method/apparatus in embodiments of the instant invention that when the method/apparatus finds a document and/or page it cannot assign or identify, it escalates the document and/or page out of the program, or automated document analysis, and displays the document to a human collaborator.
- the page is displayed on a split screen with the "heading region" of the document page amplified at the top of the screen and the entirety of the document shown in the bottom of the split screen.
- the instant inventors have determined that the identity of most documents can be determined by clues obtainable in the header region.
- Buckets is a location to store related pages during the processing involved in preferred embodiments of the instant invention. Buckets may later be correlated and classified to the operative taxonomy so that a given bucket becomes a document within the taxonomy system.
- forensic page analysis refers to a detailed extraction and mapping of the image that forms a sheet or an image of a sheet wherein this mapping is used to identify the page and/or sheet. Forensic page analysis generates a Location Diagram and Feature Vectors.
- the term "reference page” refers to the most readily identifiable document in a set of documents or pages within a Dox Package. Frequently, it is the first page of a document, but that is not required by the definition as the first page of a document may be a cover page, such as a fax cover page.
- An example of a reference page is the front page of a Form 1003.
- the "reference page” herein is the page of a document that represents the maximum logical properties or identifying properties of the document with all subsequent document members able to be classified as having affinity towards this "reference page.” This "reference page" could be, but is not necessarily, the first page of the document within a bucket or with the classified documents.
- the term "field” refers to the region of a document where specific items of information might be found.
- the individual's name is a "fact” and may also referred to herein as a "text snippet" when the fact is extracted from a field.
- fields are converted into facts by extracting the information and “scrubbing" the text output to create a value that can be utilized and/or consumed by a computer in the operation of embodiments of the instant invention.
- information fields refers to the content of the blanks on the forms, e.g., in the context of the mortgage field, the price of the property, the amount financed, the address, etc. or specific content from an unstructured document such as staled interest rate in a promissory note.
- Feature Vector refers to a manner of mapping documents wherein the relationship of keywords to fields or keywords to other keywords is mapped both as to physical distance and direction.
- the Location Diagram is a relative position map of key phrases represented in unique way by their vectors of relative distances.
- the structured files are represented in flexible structure maps called grid files.
- the term "collate” refers to the process of taking a bucket comprising a document, or a pages of a document, or sheets classified to the same taxonomy identified niche; analyzing the sheets located therein, preferably as well as all the sheets in a Dox Package, and sorting them into the correct buckets whereby all sheets belong to a document will be correctly sorted, and preferably different versions or dates of documents collected together.
- the term's definition comprises the dictionary meaning of "collates” whereby a collation occurs through a process that assembles pages in their proper numerical or logical sequence, and/or through a process examines gathered sheets in order to arrange them in the proper sequence. Collation also refers to the process of organizing Knowledge Objects into Business Objects.
- NLP refers to natural language processing, as is known to one of skill in the art.
- Image Based Classification refers to methods to classify documents using features and/or references other than text such as the visual page layout, the white-space distribution, and graphic patterns.
- the purposes of instant invention include conducting a business and making business decisions using an automated acquisition and analysis of information from a Dox Package.
- This invention thus, in part, provides:
- the documents and/or information may be converted in an XML file format such as those defined for the mortgage industry by MISMO.
- (k) it provides the ability to flexibly distribute collated documents or extracted information to a user, or sets of different documents or information to different users or decision systems using standards such as
- the instant invention provides a method/apparatus that analyzes, and collates documents, even individual versions of similar documents, preferably based on both their logical and their numerical sequence to systematically order groups of pages to enhance usability and to analyze them based on these grouped sets.
- These grouped sets are meaningful and comprehensive entities and are placed in their unique context for the specific business being supported. This collation takes place in spite of potential extreme variation in documents and in forms and the sequence of the documents or forms being input into the process.
- the method/apparatus of the instant invention is directed to a specific business, the mortgage loan business, for example.
- Jt is an object of the instant invention to provide comprehensive processes and systems that can convert relevant information from electronic images of papers and/or documents already in an electronic form to an electronic database with minimal human intervention. Further, "Knowledge Objects" are formed based on the extracted information. These Knowledge Objects may be further utilized to form "Business Objects.” The Business Objects are collations of Knowledge Objects centered on specific business requirements and can be used for subsequent decision making. An additional object of the invention is to provide a managing tool that can help in learning and configuring the overall process.
- the method/apparatus of the instant invention may collate images of sets of pages for any given type of document package (referred to herein as the "Dox Package") presented to the operator or the apparatus of the instant invention.
- documents in such a Dox Package may include images of paper documents, such as those in electronic .pdf files, native pdf files, or documents received by fax servers, for example in .tiff format.
- the instant invention is not limited to the handling of such paper documenis or images thereof.
- documents, sets of documents, pages, sets of pages, paper documents, form documents, physical pages, paper form, paper images, sheets, and the like includes documents and the like that exist in digital form, including documents, papers and forms, such as Microsoft® .doc documents and in other proprietary document formats, and the use of such are included within the scope of the present invention.
- documents may also contain embedded images, such as digital signatures or imported graphics or other documents, and likewise are included within the scope of use of the present invention.
- documents are presented or utilized following the OCR conversion of original, signed or executed, documents or a text dump of the native pdf document.
- the method/apparatus is also capable of generating its own taxonomy of buckets based on document features observed or recognized by the method/apparatus during analysis of the Dox Package.
- each page analyzed is assumed to hold a unique position within an individual document, and this page's position is determined and assigned.
- the method/apparatus initially assigns each page from the Dox Package the most logical bucket and the most appropriate position within the bucket; a page can belong to one and only one logical group.
- the position or a particular page and the sequence of pages is determined based on the page's purpose, location, readability and usability by the method/apparatus of the instant invention. After being assigned, the location of the sheet or page is preferably repeatedly re-evaluated and thus the accuracy of its position assignment, and the ultimate quality of the data, is increased.
- the method/apparatus of the instant invention provides for escalation Io a human collaborator or assistor to supplement the basic machine and expert- system-based collation.
- the level of ambiguity that triggers escalation may be preset, modified, or created during operation.
- the human collaborator can determine the identity and classification of the ambiguous document and where it should be assigned to provide clues to the method/apparatus of the instant invention.
- the present invention utilizes Location Diagram concepts and integrates multiple components including image processing, intelligent collation, feedback learning, a document classifier, a verifier, a versioning engine, an information locator, a data extractor, a data scrubber, and manual collaboration.
- the instant invention can convert representations of form documents into grid format, i.e., a text-pictorial representation. Using grid format, the instant invention can extract more and important features from the documents that then can help in formation of a Knowledge Object with very high level of accuracy and minimum human intervention.
- human review of pages within the Dox Package may only be required for one page in ten, one page in twenty, one page in thirty, one page in forty, one page in fifty, one page in sixty, one page in seventy, one page in eighty, one page in ninety, one page in one hundred, or less.
- the need for segregation, organization and collation of documents in the Dox Package arises from a number of reasons: (1) checking for completeness of the Dox Package, i.e., whether all documents required, necessary or desirable to those entities having an interest in the information contained in the Dox Package, are present in the Dox Package; (2) legal aspects of the information contained within the Dox Package; (3) business aspects of the information within the Dox Package, (4) extracting data from a large number of "hard-cop/'-only documents or images thereof, which may only be possible from a 'representative number' dur to time or money constraints, (5) requiring rapid and inexpensive access to the data contained in the documents for analysis; and (6) having available or distributing documents or sets of documents in a segregated manner based on type of document or other criteria; and (7) making decisions based on the extracted data, including compilations, aggregations, and analyzed or processed sets of such data, optionally with an automated rules engine.
- the instant invention provides a comprehensive process and system which can convert information on papers or images to an analyzed and organized electronic form where it can be used for business decision-making.
- the present invention solves the problems of sorting into versions, sequencing and collating documents and extracting information for specific industries.
- one object of the instant invention is to provide users with separated, collated and sequenced documents.
- Users of the instant invention provide the method/apparatus a document set obtained in their course of business, a 'Dox Package,' which is then collated and analyzed to meet their business requirements.
- all documents are provided at once in one location, although such documents may be provided at different times and from different locations.
- a feature of (he invention is that paper documents that do not have all the data contained therein in a segregated digital form are readily used with the instant invention.
- This invention in some preferred embodiments, comprises a comprehensively automated process that can convert data from documents in paper form to electronic form without with little, if any, human intervention.
- the instant invention may collate and classify documents based on Location Diagrams, which are based on Feature Vectors and connectivity/relationships among them. Further, the engine used in the instant invention can locate and extracts information from documents based on these Location Diagrams with additional scrubbing.
- the product is equipped with learners, which work on Location Diagram distance maximization within and across the document classes to optimize results, a "studio” (user-friendly interface) and a warehouse for storage and making data available as required by the operator of the instant invention or others designated by the operator.
- This invention may use methods of solving Location Diagrams based on simultaneous equation- and weight-based confidence measurements.
- the invention may provide significant benefit to all industries that handle sets of documents, and in particular, large, disparate sets of documents, by accelerating and improving accuracy to current decision-making process when compared to existing and traditional - methods/technologies.
- the instant invention in some embodiments, provides a method/apparatus that collates and analyzes a set of documents.
- the apparatus automatically employs various algorithms to identify groups or logical units of documents. These algorithms work to complement one another to yield higher quality results.
- the method/apparatus of the instant invention utilizes and takes into account discontinuities, for example, a page break in the middle of a sentence, to assemble pages of a document. Each of these logical units is a complete document identified as to its business identification and mapping to location within the taxonomy.
- the invention's method/apparatus preferably measures Tclatedncss among various pages; to accomplish this the method/apparatus works on the principle of a reference page.
- a "reference page” is a page that represents the maximum or near-maximum logical properties of a particular document, and thus all the subsequent document members have affinity towards this reference page.
- a reference page frequently is, but is not required to be, the first page of a given document.
- the logical sequence of a Dox Package is related to its purpose, location, readability and usability. Grouping and collating using the principles of the instant invention is concerned with completeness, usability, integrity, and unique occurrence.
- the classification and collation unit as used in the instant invention in a preferred embodiment has an Image Based Classifier, a set of text based classifiers, versioning engine, an intelligent collation engine and a verifier.
- the text based classifier preferably has a set of classification engines and each classification engine confidence is prioritized based on its strengths in handling particular (ypes of documents as will be determined by the particular application and recognized by one of skill in the art operating the instant invention.
- the reference page identification method in one embodiment uses a hybrid approach where an affinity determination method is used in connection with an input dictionary, but can also provide feedback to enhance and/or enhance the input dictionary.
- This dictionary preferably not only provides a list of words but also gives quantitative relevance of words and phrases with reference to each class of document. Keywords and keyphrases have a high affinity towards a given document.
- word 'Wl ' is defined as having a very high chance of occurrence in document 'Dl ' (e.g., the word 'interest' ('Wl ') in a mortgage note ("Dl ')) then, according to the uses and principles of the instant invention, the word 'Wl ' has high affinity towards document 'Dl .
- This affinity may be determined using Bayesian analysis and is represented as a probability or a conditional probability.
- Other Feature Vectors such as font size and type may also be considered in determining the affinity of a page to the reference page of document being examined. There is no limit to the number of Feature Vectors that might be considered for affinity analysis.
- the method/apparatus employs a multi-level approach to identify documents.
- the first pass, or Level-1 approach identifies some of the reference pages efficiently and quickly.
- Level-1 analysis may identify some reference pages along with their respective classes.
- attempts arc made to identify classes for the remaining pages.
- Level-1 uses various statistical algorithms, e.g., algorithms based on SVM and Bayesian.
- the Level-1 reference page identifier is integrated with multi-algorithm classifier which selects the best of set of algorithms based on input data.
- the classifier takes advantage of various methods like word phrase frequency, Bayesian analysis, and SVM, but is not limited to these methods and has the capability to give priority and higher weight to the most suitable method to be used for the given document for maximum accuracy and usability.
- Location Diagrams and Feature Vectors are neither required nor generated.
- Documents identified by Level-1 algorithms as ambiguous or as having affinity for more than one taxonomy class proceed to Level-2 analysis.
- the verifier used in the Level-2 analysis is preferably capable of resolving the ambiguous document classes leftover from the Level-1 analysis.
- the instant invention atso can resolve and relate documents belonging to multi-class families and documents that are within families or a group of classes that are similar.
- the verifier produces the final identification in these multi-class scenarios, using a combination of voting and critical- feature-based class verification.
- Level-2 analysis all the documents that are unable to be classified in the Level- 1 analysis are processed using critical-feature-bascd verification approach. In other embodiments, all documents are processed using the Level-2 analysis.
- the Location Diagram Map approach used in the Level-2 analysis in some preferred embodiments of the instant invention provides the required discrimination and accuracy to handle ambiguous documents and correctly classify or collate them. In preferred embodiments, this Level-2 reference page identifier uses critical-feature- based verification and voting, along with the verification algorithm and is referred to as the "verifier.”
- the collation process provides for documents to be given a logical and/or numerical sequence.
- a Dox Package is collated with reference to a prescribed or developed taxonomy where the taxonomy classes are characteristic of the industry, (e.g., industry standards like MISMO) or required or desirable by the industry, yet may be adjusted by the operator/user of the instant invention.
- Each document and/or page within the Dox Package is mapped to a class according to the taxonomy.
- the method/apparatus of the instant invention classifies these documents and collates sets of pages for industry standard taxonomy like MISMO, or any given taxonomy.
- a further feature in some embodiments of the instant invention is that the method/apparatus of the invention is also capable of generating its own taxonomy based on document features it observes. The overall method assigns most logical document structure based on the taxonomy and most appropriate position within each document for each page.
- the instant invention may thus assign meanings to the documents and put them in their proper business context by the use of the separation, sequencing and collating methods described above.
- Each document, group, or subset formed within the Dox Package is based on the document's, group's or subset's use in the relevant business.
- the system has human collaboration along with its basic machine learning and expert system based collation.
- the method/apparatus of the instant invention is also equipped with a fact extractor for use with the pages, documents or sets of documents in the Dox Package.
- This fact extraction capability provides for locating and extracting the information/fields required for various business/compliance requirements and transforms the information contained therein to facts or data that can be subject to further use or manipulation.
- the fact extractor is also equipped with weight-based confidence measurement. The fact extractor enables, in part, facts of all types and coming in various forms in the original documents to be accessed, extracted and/or manipulated.
- the fact extraction feature of the instant invention allows for human collaboration for exceptional or problematic documents, although such human intervention is not required.
- the instant invention can handle all types of fields, e.g., OMR, tables, descriptions, numbers and the like, that will be known to those of skill in the art, depending on the particular business application.
- the decision system lhat is optionally used as part of the instant invention provides logical decisions based on this information obtained or extracted and the relevant business context.
- the preferable need-based human collaboration built into the system makes it possible to extract information and/or data from fields with a very high level of accuracy and coverage.
- the instant invention also provides a decision-engineering-framework specific to the business application to organize and utilize the extracted information.
- the information extracted from a Dox Package is preferably presented in a usable format, such as a spreadsheet or XML tag file format.
- automated decisions may be made on the information obtained by an automated rules engine such as Microsoft BizTalk, ILOG jrules, etc.
- an appraisal report regarding a piece of property (most of which are created as PDF files, if they are available electronically at all) and extract the information (including unstructured information), to create an XML output.
- This output can be used for a variety of purposes such as it may be furnished to a company that evaluates and scores the accuracy risk of the appraiser's information, to generate a report similar to an AVM to a mortgage banker for a business decision.
- the instant invention may convert the information from an appraisal into electronic data over 100 times as fast as a human operator and with better accuracy..
- the instant invention can extract information automatically from a credit report. This information may be furnished, for example, to a mortgage lender for their risk assessment process.
- the instant invention provides for collation of all the pages, documents, or sets of documents within a Dox Package into a taxonomy classification to meet the business needs of the operator and/or a particular industry.
- Virtually any Dox Package from any industry may be analyzed by preferred embodiments of the instant invention.
- the collated documents are mapped as to a taxonomy such as MISMO or any other industry-specific or user-specific taxonomy.
- information is extracted from this Dox Package. This information is scrubbed and transformed into discrete data and/or facts. The facts and its related information is used to form an information matrix called a Knowledge Object.
- the Knowledge Objects are transformed in a particular or required business context to create Business Objects.
- the Business Objects are then used for business decision-making.
- the instant invention therefore facilitates extraction of critical information for businesses from the documents, and provides for manipulation, compilation, analyzing and/or access to the facts or data or creation of transaction sets that comply with the MISMO SMART DocsTM standard and/or other custom XML tag file formats.
- the instant invention in preferred embodiments, provides for an automated method with exception- based human collaboration ("escalation") to collate with increased speed and accuracy;
- the instant invention in preferred embodiments, provides superior accuracy and quantifiable measures for accuracy
- the instant invention in preferred embodiments, is the only comprehensive collation solution which can collate pages, documents, or sets of documents identified by revision numbers, for business decision making purposes;
- the instant invention in preferred embodiments, is not limited to document separation by boundary detection algorithms
- the instant invention in preferred embodiments, provides for the mapping of documents and document images to a JvI ISM O taxonomy, as well as other industry standard and custom taxonomies;
- the instant invention in preferred embodiments, locates and extracts information from documents and document images sorted into buckets with a high degree of accuracy
- the instant invention in preferred embodiments, provides, among others, the features of intelligent scrubbing and fact conversion and/or other data manipulation features; the fact conversion converts extracted information into data or facts that offer value to businesses and provide direct input into an automated rules engine using custom or industry standard XML formats such as those specified by
- the instant invention in preferred embodiments, provides an automated learner which can accommodate and incorporate new document types, and the intelligence to deal with variations in the number and type of documents and field locations;
- the instant invention in preferred embodiments, can incrementally learn to adapt to changes in the patterns between and/or within documents;
- the instant invention in preferred embodiments, provides a Location Diagram-based extraction for accurate extractions in case of slippage, variations and changes in format;
- the instant invention in preferred embodiments, features in some embodiments a collation confidence matrix to be able to assess the confidence level of the method or algorithm, plus the instant invention, in preferred embodiments, may effectively use all clues gathered during all phases of document processing and analysis to validate the accuracy of the result.
- business decisions such as whether to invest in a bundle of loans on the secondary market, may be based on extracted information from a large number of the associated Dox Packages, or a majority of the associated Dox Packages, or almost all of the associated Dox Packages.
- One of the advantages of embodiments of instant invention is the number of discrete pages it can analyze.
- the instant invention may handle more pages and more diverse pages than what was present in the art.
- 2, 3, 4, 5, 6, 7, 8, 9, 10, 13, 15, 18, 20, 25, 30, 35, 40, or more pages may be analyzed in on Dox Package.
- 2, 3, 4, 5, 6, 7, 8, 9, 10, 13, 15, 18, 20, 25, 30, 35, 40, or more document types may be analyzed in on Dox Package.
- the systems and methods described herein provide the capability to ensure that the right information is always available to the right people at the right time.
- outsourcing and offshoring typically require a paper-based system
- paper-based system can be converted to images using systems and methods described herein to unlock data that needs to be accessible to make decisions.
- trailing docs are automatically sent to the correct electronic folders, bypassing mailroom delays, and even brokers offices are made virtually paperless. Tasks can be handles as easily offshore as onshore, if such outsourcing fits into the lender's plans.
- the instant invention is, in preferred embodiments, a process and system for separating, organizing and retrieving information from various documents, for example from a Dox Package.
- the system preferably employs a collator, a classifier, an extractor, a scrubber, a verifier, a version engine, a voting engine, a transformer for creation of Knowledge Objects and Business Objects, a decision engine and a learner for classification and extraction.
- FIG. 6 An exemplary embodiment of the invention is illustrated by Figure 6 depicting a deployment method to execute a Web-based method for intelligent paperless document management.
- any device capable of reaching the internet can be used for accessing a paperless workspace of the subject invention.
- the paperless workspace can be represented to the device through a web browser as a Rich Internet Application 603 or through an installed software application 602 local to the device that accesses the paperless workspace securely via the Internet.
- the applications that represent the paperless workspace can communicate securely through the Intelligent Paperless Document Management Network (IPDM Network 604).
- IPDM Network 604 can act as the gateway and hub to route all common services and host the engines required by the systems and methods described herein.
- Domains 605 and Client Sites 606 are logical subsets of the full functionality available via the systems and methods described herein.
- the Domain A in 605 may be a private labeled sales organization that markets and fulfills demand to certain industries, such as "LoanKatalyst” primarily caters to the industry comprised of US Mortgage Bankers/Brokers.
- the LoaniCatalyst Domain (Domain ⁇ ) is a sales organization that has used the systems and methods described within to pre-define capabilities for the attached Client Sites that primarily house "Mortgage Loan Folders.”
- the Client Site 606 comprises the entirety of secure data and configurations for each mutually exclusive Intelligent Paperless Document Management for an organization and/or user.
- Client Sites A l, A2, and A2 in 606 within the LoanKatalyst Domain may be restricted to a pre-defined sub-set of capabilities outlined by the systems and methods described herein, whereas Client Sites within a different domain may employ a different sub-set of capabilities.
- Domains 605 described herein allow sales organizations to constantly provide the right type of functionality to the right people and organizations.
- the systems and methods described herein require at least one configured IPDM Network 604, at least one configured Domain 605, and/or at least one configured Client Site 606.
- the Web-based Rich Internet Application representation of the Paperless Workspace 603 can be critical to providing an authorized end-user a user experience that enhances the decision to "go paperless” by enabling the user to have a digital paper experience as friendly and accessible as using paper.
- application development tools such as AJAX, Adobe Flex, etc
- the systems and methods described herein can be deployed either on the desktop through compatible operating system extensions such as Adobe AIR, or through Web-based browser plug-ins such as Adobe Shockwave Flash.
- the systems and method described use the deployment of a Paperless Workspaces Application as a Rich Internet Application.
- FIG. 7 An exemplary embodiment of the invention is illustrated by Figure 7 depicting the configuration management and administration of a Web-based method for Intelligent Paperless Document Management.
- a flexible configuration and administration chain having various modules such as the IPDM Network 701, Domains 710, and Client Sites 720 and 730 can be deployed to form the foundation of administration for the systems and methods described herein.
- step 701 Authorized Network Administrators (at the right side of Figure 7) can be used to configure and administrate the IPDM Network.
- the IPDM Network Administration 701 can include at least one of the following components: document classification and extraction management 702, paperless workspace application deployment 703, Site-To-Site delivery management 704. secure application program interface management commonly referred to as "APIs" 70S, management of custom interfaces from the network to export documents and data to third party data and document systems 706, management for creation and administration of reseller organizations 707, as well as the management for creation and administration of Domains 708.
- APIs application program interface management
- step 702 the extension of auto-classification (automatic indexing) and data extraction from submissions of organized or unorganized digital paper can be used to create the foundation for the deployment of Intelligent Paperless Document Management, where such Intelligent Paperless Document Management can stand up to the real world needs required to eliminate paper from an organization.
- Step 702 can use document classification (automatic indexing) and/or extraction engines described herein or any other commercial available auto-indexing/extraction engines. These engines can be coupled with respective libraries of document dictionaries depending on the industry of choice. Document dictionaries can be defined where a series of learned documents are grouped together to efficiently solve a business problem identified industry by industry.
- the "US Residential Mortgage Loan Document Dictionary" used by the LoanKatalyst Domain provides the capability to create an extendable group of learned documents that can be used in the auto-classification of submitted documents through an Inbox that is assigned to a LoanKatalyst Client Site (such as Client Site Al in Figure 6 and a Client Site in 720 here).
- Each Client Site can take the advantage of using an auto-classification (automatic indexing) engine described herein, or any other commercial classification/extraction engine available for license.
- step 703 at least one paperless workplace app deployment can enable and manage authorized end users to access digitally represented paperless workspace using any device that can access the Internet through these applications, ⁇ n step 703, multiple applications can be devised and deployed to provide authorized end-users with a rich interface that can model and enhance a user experience similar to actual paper. Due to emphasis on end-user experience, different applications may be needed for different types of users to provide a tailor-fit paperless workspace experience to an intended user at every time.
- Step 704 shows the configuration and management of Site-To-Site delivery of documents and/or data among various Client Sites.
- Client Sites can be provided with capabilities and administrative tools required to allow delivery of documents and/or data from other Client Sites to their Client Site.
- a Mortgage Lender who uses a Client Site Al on the IPDM Network, can receive documents of a new loan submission from a Mortgage Broker, who also uses a Client Site A2 on the IPDM Network, using the Site-To-Site delivery in step 704.
- the IPDM Network can provide all the necessary services required to deliver authorized groups of documents & data through the network securely without requiring the digital package to travel outside of the network.
- Step 705 shows the management of application program interfaces (APIs) that allow authorized third party developers to create interfaces from third party systems to the systems and methods described herein.
- APIs application program interfaces
- Step 706 shows the management of custom interfaces developed specifically using third party APIs or industry standards to enable the link between the systems and methods described and third party systems based on business rules and authorizations specified at the Domain and Client Site levels.
- Step 707 describes the management capabilities for the Authorized IPDM Network Administrators (at the right side of Figure 7) to create and administer distribution relationships called "Resellers" that may distribute Client Sites under certain specific Domain(s).
- Step 708 describes the management capabilities for the Authorized IPDM Network Administrators (at the right side of Figure 7) to create and administer Domains for extending the flexibility of IPDM Network systems and methods down through a Domain to be used by the Client Sites.
- the Domain can be created to extend and make available different types of functionality based on the needs and requests from Domain Administrators.
- Client Sites can be shown with the LoanKatalyst Domain labeling.
- the LoanKatalyst Domain provides the capability to extend the flexibility of the systems and methods described herein to create a user- friendly way to create a paperless mortgage loan for use by all interested parties in the US Mortgage Industry.
- Typical administration tasks can include at least one of the following components: the extension of Document Dictionaries to link to Inboxes, Folder Types, and Workflow Queues, as well as the exporting of documents and/or data from the system to third party mortgage data or document systems.
- the systems and methods described herein can provide extension by Domains to an easy to use, yet incredibly advanced and paper-like business method of creating a paperless transaction and paperless workplace in any industry applied. (00182) E-Signatures are quickly becoming a viable method for execution of documents within the United
- E-Signature protocols for the IPDM Network are depicted in step 709.
- E-Signatures requested by subscribing Client Sites at the Online Client Site level can extend the IPDM Network's E-Signature systems and methods. Recipients of deliveries requiring E-Signature can access the Document Pickup and E-Sign Portal (as shown in Step 821 of Figure 8) through a link.
- the Recipients can create an E-Signature account on the IPDM Network. Once authorized E-Signatory account holders, Recipients can download and install an application that works in conjunction with the Document Pickup and E-Sign Portal to validate the identity of E-Signatories against the documents required for E-Signature. In some embodiments, E- Signatories may have future access to download or print copies of the documents provided for E- Signature. Once the E-Signatures are collected, the information can be vaulted and encrypted using industry electronic signature standards on the IPDM Network and the recorded transaction and documents can be accessible to the Online Client Site. Delivery of documents bearing E-Signature data can be handled on an industry by industry standard.
- Step 710 shows the management on the Domain Level by the Authorbxd Domain Administrators (at the right side of Figure 7).
- Authorized Domain Administrators can be provided with the ability to use at least one of the following available interfaces: the management of private labeling per Domain for Resellers 711 , the creating of new Client Sites 712, and the management of the accounts of previously setup Client Sites 713.
- the Authorized Domain Administrators can be provided with necessary tools for putting private labels onto the Client Site applications that will attach to that Domain (Step 711).
- Such management tools can include at least one of the following capabilities: the ability to change logos, color schemes, automated messages, and other branding opportunities. Due to the enormous flexibility of the systems and methods described herein to deliver Intelligent Paperless Document Management, distributions of Client Sites can be maximized by having Domains that focus specifically on extending the capabilities to focused verticals of expertise by distribution agent or Reseller, using the systems and methods in such a way as to make the smoothest transition to paperless for each Client Site without requiring intimate technical knowledge.
- Step 712 shows that Authorized Domain Administrators (at the right side of Figure 7) can directly create new Client Sites or enable authenticated APIs to have Client Sites sign up through a Web-based form to auto-create a new Client Site attached to the Domain.
- a new Client Site can be automatically created for the LoanKatalyst Domain with a click of a button. Due to the complete prior configuration of the IPDM Network and that of the Domain, creation of Client Sites attached to the LoanKatalyst Domain can be automatically generated within a matter of seconds using the systems and methods described herein. This is advantageous to the distribution to users enhancing the perception of an easy-to-deploy paperless workplace.
- Authorized Domain Administrators can be provided with management tools to manage the accounts of the Client Sites, as shown in Step 713.
- management tools include, but are not limited to, managing the status of Client Sites, canceling a Client Site from the Domain, and modifying the subscriptions to services/methods/apparatus that are made available to the Client Site through the Domain.
- step 720 at the Client Site level, Authorized Client Site Account Administrators (at the right side of Figure 7) can manage each Client Account and their subscriptions separately through a Client Site Subscription Management portal to services offered by the Intelligent Paperless Document Management systems and methods defined herein.
- the Client Site Subscription Management portal 720 can use at least one of the following management tools: Client Site account management 721, and management of subscriptions to Inboxes 722, Folder Types 723, Workflow Queues 724, and Document Dictionaries 725. Changes to subscriptions within Client Site Subscription Management 720 can yield direct changes to further administer on the Online Client Site.
- the Client Site Account Management can provide means for Authorized Client Site Subscription Administrators (at the right side of Figure 7) to manage their account status, account information, and/or their list of Authorized Client Site Subscription Administrators.
- the Client Site can be provided with the ability to extend its functionality by subscribing to additional services offered through the Domain.
- One such extension is the capability of expanding the list of Inboxes that allow intake of documents and data from both third parties and authorized users.
- a Client Site must have at least one Inbox that is configured to allow for the transmission of digital data through a designated Fax Number, Email address, SFTP site, HTTP adaptor, etc.
- This method describes a publically accessible Inbox which can be further expanded through the Inboxes Subscriptions Manager 722.
- a Client Site may wish to have multiple publically accessible Inboxes to extend its business practices.
- a Client Site on the LoanfCatalyst Domain can provide at least one following publically accessible inboxes: Company Inbox, New Retail Loans Inbox, New Wholesale Loans Inbox, and an Employee Inbox for every employee.
- Publically accessible Inboxes may require the IPDM Network to assign a list of addresses such as Fax Numbers, Email Addresses, SFTP and HTTP locations. These locations can be publically accessible to transmissions of documents and data to those addresses to be routed to the subscribed Inboxes.
- Folder Types are a type of extendable container for documents with associated, configurable and searchable meta-data attributes and document properties.
- the LoanKatalyst Domain can make accessible a "Mortgage Loan Folder" type that has a predefined list of attributes specific to what a mortgage industry professional may use to describe, search for, and organize a paper-based mortgage loan folder such as text fields entitled "Borrower Name,” “Loan Number,” “Property Address,” “Loan Officer,” “Processor,” “Lien Position,” etc.
- a Client Site can provide a user the ability to create a New Folder by creating an instance of a Folder Type for which they subscribe, which would then extend the attributes and document properties of the Folder Type to that instance of the new Folder.
- the Folder Type repository can have additional modifiable properties that may be specifically associated to any documents that reside "within” an instance of that Folder Type.
- each Domain can be provided with the ability to extend multiple pre-defined Folder Types for which the Client Sites may choose to subscribe.
- a Domain can be provided with the ability to allow the Client Site the ultimate flexibility to create their own Folder Types, and allow the Client Site to subscribe to a number of those custom Folder Types in step 723.
- Step 724 shows the Workflow Queue systems where the methods described herein can be extended to the Client Site through subscription by Client Sites that wish to deploy workflow through Intelligent Paperless Document Management
- a Client Site may merely subscribe to them and administer them at the Client Site level.
- Client Sites can be provided with capability to manage their subscriptions to document dictionaries of learned documents as well as to submit requests to the TPDM Network to leam and incorporate new documents into a subscribed document dictionary that are custom to that Client Site.
- Step 72S can truly bring the power of auto-classification and extraction to the fingertips of business administrators.
- a Client Site on the LoanKatalyst Domain may have an internal document named "New Wholesale Loan Checklist" which is a PDF document that is internal to that user organization and is filled-out by hand with data using Adobe Acrobat for each loan that comes into the organization.
- Such a Client Site may prefer that the document be auto-classified and may also wish for the data that is input to the PDF be extracted in a certain format.
- Step 725 can allow the Client Site to electronically submit samples of the document and data that the Client Site wishes to be learned, and subscribe to having that document/data added as a custom learned document to one of their subscribed document dictionaries.
- Step 726 may use Step 726 to manage their subscription to E-Signatures. If a Client Site would like to deploy E-Signature methods within their Online Client Site, they may manage their subscription to E-Signatures using Step 726.
- the Authorized Client Site Administrators can interact directly with the Online Client Site Administration Portal to manage and configure all of the features of the Online Client Site.
- administration abilities can include at least one of the following: Folder Type administration 731, Online Storage administration 732, Archived Storage administration 733, Role-based privileges administration 734, User administration 735, Inboxes administration 736, Workflow Queues administration 737, Address Book administration 738, Document Dictionaries administration 739, Document Stacking Order administration 740, and Site-to-Site Delivery administration 741.
- the Online Site Administration methods can provide a targeted, user friendly environment to deploy very advanced security and access functionality.
- the methods can be designed in such a way to make the Administration of Intelligent Paperless Document Management accessible and understandable to business level users without the support of specific technical resources or knowledge.
- Client Site Administrators can be those end users of a Client Site that have been given Administrative access to the entirety of the Online Client Site Administration Portal or any subset of Administrative methods. These end users can be different from those described in Step 720 for accessing Client Site Subscription Management.
- the first step in administering an Online Client Site can be to administer the Folder Types that may be setup using the Online Client Site.
- Step 731 can be an extension of Folder Types administration into the Online Client Site.
- Client Sites that have subscribed to a single Folder Type through a domain, or have subscribed to the capability to create Custom Folder Types can be offered Online Administration of these Folder Types.
- Authorized Client Site Administrators can be provided with the ability to customize Folder Types to enhance the searchabiiity and usability of Folder instances of those Folder Types.
- Folder Types have properties that may be turned on and off to enhance the security model of the organization.
- Kolder Types have searchable attributes that can be created and defined to enhance the usability of the Folder instances created.
- Folder Types can have configurable document attributes that can be attached to documents that are filed into an instance of the Folder Type.
- Folder Types can also have configurable named default Document Sets that can be setup to enhance security and usability.
- the second step in administrating an Online Client Site can be to setup Online Storage repositories to house Folders.
- Administration of Online Storage repositories of documents can the method described by Step 732.
- Folders can reside in Online Storage repositories that can be configured per Domain in a hierarchy that best represents the industry that the Domain services.
- a Client Site can have access to Mortgage Loan Folders that arc stored in online Cabinets, whereby the online Cabinets reside hierarchically in online Offices.
- Administration of the Offices and Cabinets can create online repositories to store online Folders. Offices and Cabinets can be created, named, and grouped in such a way that best supports the individual Client Site.
- a LoanKatalyst Client Site may have an Office named "Retail Sales - San Diego, CA Office” holding Cabinets named after die individual Sales representatives.
- the possibilities for setups of the Offices & Cabinets can be endless per Client Site per Domain per Industry in the hierarchical method described.
- Administration of the online storage can also allow for the renaming and deletion of repositories to suit the needs of the Client Site.
- Step 733 shows the Archived Storage administration.
- a Folder or group of Folders can be moved to an archived facility or converted to a non-indexed medium and can be held in boxes in physical warehouse storage to reduce the costs and clutter of keeping the paper folders indexed and within reach.
- Such a need also exists in the paperless world, as the creation of online folders depending on the types of folders and industries they are created in may eventually clutter up the repository and the costs of keeping the folders online would outweigh the benefit Therefore, the extension of Archived Storage can be administrated through the creation of repositories.
- the Archived Storage administration method can allow for the creation of Archive Warehouses and Boxes per Folder Type to keep Folders that are required to be archived. These Archive Warehouses can hierarchically house archive Boxes, which contain archived Folders.
- Step 734 depicts the administration method for Role-based privileges.
- a Client Site can have different levels of functional capability.
- An Authorized Client Site Administrator can create an unlimited number of Roles for which they may allow or deny any number of Client Site privileges to each Role.
- Another way to describe Roles is that they are groups of pre-defined user privileges that may be assigned to Users to enable ease of privilege administration.
- Conveyed privileges can allow a user assigned to a Role to enact events on the Online Client Site.
- Examples of such privileges can include, but not limited to, creation of folders, editing folder attributes, moving folders from one Office/Cabinet to another, archiving folders, deleting folders, editing documents, deleting documents, delivering documents by fax, delivering document by email, adding notes to documents/folders, deleting notes on documents/folders, editing document attributes, and more.
- the systems and methods described provide at least one privilege listed above.
- Step 735 depicts User Administration, which can be the actual method to create and administrate a list of End-Users who may be granted access to portions of paperless workspace application software digitally representing the Online Client Site.
- Enabled Users can be given the authority to use their private login credentials to access the Online Client Site. Disabled Users can be denied access to the Online Client Site.
- Enabled Users may have their password reset by a Client Site Administrator at any time, and must follow login/password rules as defined in the system by Client Site Administrators.
- Enabled Users may be given access to different Folder Types. Users may be assigned a Role that governs the privileges the User has been conveyed when accessing a Folder of certain Folder Type. Users may be assigned different roles to each Folder Type they are granted access.
- Storage access may be granted by giving the User access to All Offices, or a mix of single or multiple Offices and/or single or multiple Cabinets. If a User is given access to a cabinet, the User can have access the Folders of a Folder Type that they have permitted that reside in the Cabinet(s) assigned. If a User is given Office level access or All Offices access, then the User can have access to all Folders of a permitted Folder Type across all the Cabinets in that Office or across All Offices, respectively. If Queues are extended through the Client Site Subscriptions, the User may be assigned to specific Queues to perform work in those Queues. Furthermore, Users may be granted access to shared Address Books. 100203) Inboxes administration for the Online Client Site is depicted in Step 736.
- Public Inboxes can be inboxes deliverable to by publically addressing the Inbox via Fax or Email addresses, as well as giving access to a third parties via SFTP/HTTPS/APIs.
- Public Inboxes can land in the Client Site as the result of subscription to the Inboxes from the Domain/IPDM Network.
- Inboxes may be assigned to a Document Dictionary subscribed to by the Online Client Site for which the learned documents will comprise the extent of auto-classification and extraction for transmissions through that Inbox.
- Private Inboxes can be inboxes deliverable to only by Authorized Users of the Online Clieni Site or through authorized Sitc-To-Site delivery methods. The Authorized Client Site Administrators may create, edit, or delete Private Inboxes.
- Inbox Managers may be setup to subscribe to Inbox Alerts to an external email client, SMS, or workspace alert based on the type of event that happens in an Inbox, such as, but not limited to, being alerted upon the receipt of new Mail Items in the Inbox and being alerted if Mail Items in an Inbox are overdue for filing.
- a Client Sites that wishes to deploy Inboxes to each of their 25 employees may setup 25 Inboxes (either Public or Private), name the Iuboxes in a way to designate that the Inbox is for Employee X, and add Employee X as the Inbox Manager.
- Inboxes either Public or Private
- name the Iuboxes in a way to designate that the Inbox is for Employee X
- Employee X add Employee X as the Inbox Manager.
- Users with Public Inboxes can eliminate having sensitive documents sent solely via unsecured email to their email clients, or deploying or subscribing to third party fax-to-email servers, and can instead take advantage of Fax/SFfP/Secured Email/HTTPS/Secured Upload directly to their Inbox, and have the capability to file those documents and data securely in a repository.
- Administration of this very complex deployment of Public and Private Inboxes can be easy enough for business users to deploy without requiring specific technical knowledge.
- Step 737 Paperless workflow can be as accessible and easy to use as the paper-based workflow of dropping Folders in a physical paper tray, while also providing all the benefits offered by a paperless system. Administering and deploying workflow in a paperless environment previously requires a tremendous amount of technical knowledge. In some embodiments, Workflow Queue administration for the Online Client Site of the subject invention does not require such technical knowledge.
- the Authorized Client Site Administrators may create, edit any aspect of, or delete a Queue.
- a Queue may have one or more Queue Managers.
- a Queue may be associated with a Folder Type.
- a Queue may have one or more Work Tasks defined within.
- a Queue may have one or more automated Alerts defined.
- a Queue may have one or more automated or manual processing rules assigned.
- a Queue or its work tasks may have time limits for completion before generating an automated processing rule assigned.
- a Queue may be accessible by Authorized Users assigned to the Queue.
- a Queue may be stand alone, part of a series, or allow parallel processing.
- Each Shared Address book may have Address Book Managers that can be added or removed to a Shared Address Book by Authorized Client Site Administrators. Address Book Users arc Users that can be granted access to the Shared Address Books. Shared Address Books can allow the Delivery of chosen documents to pre-defined recipients.
- Shared Address Book Entries may be created and setup to deliver via any of the real world methods of delivering paper or digital paper between one human/system to another.
- a Shared Address Book Entry may be setup to deliver to a specific Fax Number, Email Address, SFTP site, HTTPS, Site-to-Site delivery, or through interface to process a hard copy to a physical address through an overnight mail carrier.
- Document Dictionaries Listed in the Online Client Site Administration Portal are the Document Dictionaries subscribed.
- Document Dictionaries can represent compiled groups of Document Taxonomies.
- Document Types listed within the Document Dictionaries may be standard or custom.
- Document Types listed within the Document Dictionaries may be part of a Category within the Document Dictionary.
- Document Types listed within the Document Dictionaries may be part of the Auto-Classification learn set.
- Document Types listed within the Document Dictionaries may be part of the Data Extraction learn set.
- Document Types listed with the Document Dictionaries may have searchable tags added to them to facilitate searching for an instance of the document within a Folder.
- Document Types may be added, edited, or deleted from the Online Client Site's Document Dictionaries using Step 739.
- Stacking Orders may contain document types of one or more Document Dictionaries. Stacking Orders may be rearranged to suit the parties necessary. Stacking Orders may be assigned to any delivery at the time of delivery, or as a default for one or more Address Book Entries.
- Site-to-Site delivery can allow any Client Site on the IPDM Network to delivery documents and/or data to any other Client Site on the IPDM Network, regardless of the Domain.
- Administration of Site-to-Site Delivery is depicted in step 741 ,
- Authorized Client Site Administrators may create, edit, disable, or delete a Site-to-Site delivery protocol between Client Sites.
- a Site-to-Site Delivery protocol may be enabled for any external Client Site.
- a Site-to-Site Delivery protocol can require two-party authentication between Client Sites.
- a Site-to-Site Delivery Protocol may be authorized by one Client Site to accept Inbound Deliveries from another Client Site through a designated Inbox.
- Client Sites may add Client Site B to a Site-to- Site inbound protocol. Client Site B may then accept the Site-to-Site protocol and may add a protocol to allow Client Site A to deliver Site-to-Site back to Client Site B. Two-way delivery can be accessible once Client Site A accepts the new Site-to-Site protocol. Client Sites may setup multiple inbound protocols for the same external Client Site to deliver through different inboxes respectively to allow for correct channeling of documents/data.
- Figure 8 depicting a Web-based method for intelligent paperless document management.
- a user First a user must set-up for accessing the secure website. After the user performs the online sign-up step for registration, an automatic e-mail disclosing details such as a password will be sent to the administrator specified during the registration step.
- the systems and methods described herein will then assign at least one intelligent inbox 806 for accepting incoming documents and ensure data is secure and accessible only to individuals approved by the administrators.
- the at least one intelligent inbox can recognize certain coversheet with routing information.
- a fax coversheet containing details describing destinations of accompanied documents can be used to facilitate automatically indexing and/or routing of the documents to the desired folder (see examples below and Figure 52).
- e-mail containing folder ID numbers in the subject line can be used to facilitate automatically indexing and/or routing of attached documents to the desired folder.
- Such intelligent inbox can accept incoming documents from various means such as fax or e-mail with various formats such as pdf, tiff, or gif files.
- IJ Incoming documents or data items can be delivered from a third party deliver via fax, e-mail, SFTP,
- documents can be directly uploaded from the Web-based offices by an "upload" button found in the document viewer.
- Authorized users can upload documents via such internet client site (see 802).
- the systems and methods described can provide site-to-site delivery from one internal client site to another internal client site if both sites have been authorized to do so (see 803).
- Documents can also be delivered from authorized APIs (see 804),
- the systems and methods described comprise a desktop tool (can be called Messenger) which can assist in scanning and uploading files (see 805 and examples below).
- the desktop tool can work with any TWAIN compliant scanner, and it will let the user navigate the Web- based offices to directly upload documents into specific folders.
- the desktop tool can also comprise a flexible print driver for printing one or more documents directly into the Web-based offices instead of printing the documents using regular printers, and then scanning and uploading the documents.
- Step 807 is for document upload and preparation.
- step 807 is similar to 201 Doc Prep in Figure 2.
- Step 807 may use modules for image capture, image enhancement, image identification, and/or optical character recognition. Any component of this exemplary embodiment can be replaced by a person of skill in the art with another commercial available component.
- any commercially available scanner or digital copier can be used to upload documents in step 801.
- the systems and methods described herein can take a file or document that has been faxed to the user, turns the file or document into at least one image and breaks the image down into various components of the file or documents. This process can be carried out without the need for inserting blank sheets as separators.
- the systems and methods described do not require pre-processing, separator sheets, or bar-coding for the purpose of indexing or classifying documents.
- Step 808 provides the use of at least one automatic or automated indexing engine.
- the systems and methods described herein provide capability to upload unorganized or organized documents, identify and classify each documents, collates pages of the documents, store the documents in a central repository within a secure FTP site, allow access of documents via a secure internet site.
- Incoming documents are automatically indexed and collated into a familiar "stack," tabbed by categories, and placed into at least one designated electronic cabinet/folder. Documents that are not recognized can be automatically labeled or labeled using a pull-down menu.
- the systems and methods described herein can auto-classify documents as they enter an organiztion's office, such as a mortgage lender or broker's office, via fax, e-mail, FTP transfer or by using the systems" print driver.
- the user interface and file management capabilities can move documents into repository as pdf documents.
- the systems and methods described provide an automatic indexing engine which recognizes over 200 document types in categories reflecting typical loan stacking protocols. Any component of the exemplary embodiment can be replaced by a person of skill in the art with another commercial available component.
- the automatically indexing in step 808 can be performed using a classification engine described in US 2007/0118391.
- the automatically indexing step 808 can be performed using another commercially available auto-indexing engine.
- Step 808 also provide the use of at least one data extraction engine which can pull data points from pages of a document.
- Data can be extracted with high precision from native pdf files.
- data extraction can be carried out from all Fannie Mae SMART Doc, because many lenders and investors continue to produce electronic loan documents in PDF format.
- Data extraction (or data capture) services are available to isolate key fields, enabling anti-fraud and other analytics at high speed for both post-closing and pre-funding applications.
- the data extraction engine of step 808 can minimize manual data reentry which is time-consuming and error-prone. If all extracted data are consistent to one another, extracted data can be stored in a searchable online electronic repository 810 in at least one specified format. If there is any inconsistency among extracted data, step 809 provides a process flagging for human intervention. In some embodiments, the human intervention can be sorting, modifying, or deleting at least one document or file.
- data extracted from an appraisal can be compared automatically to an automatic valuation model (AVM), or similar decision system, to see whether the appraisal reflects adjustments beyond lender tolerances.
- AVM automatic valuation model
- data returned from a fraud report may show multiple recordings from a borrower, and those transactions can be checked against data extracted from other investment property documents. If the information does not match, an underwriter and/or lender can be notified automatically or manually (also see step 814 and 823 below).
- Step 810 provides the step of storing indexed documents and extracted data into a central secure repository for easy access and retrieval. Instead of manually collecting and organizing files, the systems and methods described herein allow files to be uploaded into one central repository automatically. Documents are no longer spread out in multiple offices or in the hands of multiple people and all documents can be easily accessed by multiple users simultaneously via an online RlA collaborative intelligent paperless workplace 813. Most of screen shots in the examples and figures are taken from an exemplary online RIA collaborative intelligent paperless workplace 813.
- the indexed documents can be organized into electronic folders in pre-determined electronic cabinets. Each organized document can be searchable and can be viewed by any user allowed for access within minutes after the receipt of the incoming documents.
- access to the central repository 810 from a desktop tool such as Messenger 805 can be authorized.
- users can view documents in the central repository 810 without going through the online collaborative intelligent paperless workplace 813.
- Step 811 provides the use of searchable Long-Term Storage. Although files remain online in the Web- based offices, specific cabinets can be designated for long-term storage purpose only. Files under long- term storage can be easily retrieved such as using universal folder searches for purpose such as compliance checks. Searches can be performed using various fields used when the folders were created previously.
- the systems and methods described herein allow documents to be delivered electronically and/or archived to CD/DVD's or to users' local computers. Users can save selected documents onto hard drives of local desktop or laptop computers and/or onto one or more removable storage media such as CD-ROM or DVD-ROM.
- an "export" feature of the systems and methods described can allow users to export documents into local personal computers for local storage.
- Step 812 provides the use of at least one online client site administration portal.
- An exemplary administration portal is illustrated in Figure 7.
- Step 813 provides the use of online RIA collaborative intelligent paperless workplace which allows many permitted users to view and take actions on indexed files or documents simultaneously, with full security. Once documents are uploaded and indexed, everyone with authorized access to the documents can instantly view any of the documents via internet.
- the online RlA collaborative intelligent paperless workplace 813 functions as "software as a service" over the Internet. In some embodiments, the only types of software users need to use the systems and methods described are Flash® player and Acrobat Reader® by Adobe Systems Inc. San Jose, CA.
- the user interface of the online RIA collaborative intelligent paperless workplace 813 can take the viewers directly to the documents, extracted data, and/or information desired to make decisions.
- a loan processor can click to open the documents and view them on screen. Notes can be recorded as needed right onscreen rather than on sticky notes and scrap papers.
- the loan processor can simply drag the icon into the underwriter's cabinet.
- the underwriter has completed their tasks, the loan can be dragged and dropped back either in loan processing or in the closing department's cabinet where it is queued up and waiting.
- Such workflow queues can be viewed and monitored as provided in step 814 using an online collaborative interface. Either the online collaborative interface for workflow queues 814 or the decision system/database of record 823 can be linked to the data extraction engine 808 as described earlier.
- the linkage between the data extraction engine 808 and the collaborative interface for workflow queues 814 can provide automatic detections of important tasks or automatic notifications of important information.
- the linkage between the data extraction engine 808 and the decision system/database of record 823 can provide certain automatic decision making to facilitate business decisions.
- Access options can include universal access, access to specific offices only, or even specific cabinets within offices. Permissions to access can be limited using present roles or by creating custom roles.
- Once user roles and access privileges are determined, the systems described will e-mail users with login information for instant access. The securitization function of the systems described can allow a user to correct and/or notify errors already viewed by other users.
- the systems and methods described herein can be integrated with at least one desk-based system such as mortgage loan origination software (LOS), automatic underwriting systems (AUS), lender underwriting systems (LUS), automate mortgage compliance software, or other electronic loan software.
- LOS mortgage loan origination software
- AUS automatic underwriting systems
- LCS lender underwriting systems
- automate mortgage compliance software or other electronic loan software.
- the systems and methods described can determine specific documents required by a particular person or party and automatically import or send documents into the person or party's system.
- the systems and methods described can also be integrated with other web services.
- the systems and methods described herein allow multiple people to access documents simultaneously from difference locations through password-protected access via internet.
- the systems and methods described organize electronic cabinets by workflow steps and use a "move" feature allowing users to send documents from cabinets to cabinets. Several people can have access to the same file or document simultaneously as it can be viewed on computer screens at the same time.
- Step 814 can also provide at least one Business Rule Manager based on at least one Workflow Rules.
- Step 815 provides the use of online collaborative folders, where each folder comprises a different metadata.
- Step 816 provides an outbound delivery center for document delivery, where files can be sent via various means such as secure e-mail, hard-copy fax, direct upload, or overnight delivery. Delivery can be carried out among anywhere in the world with internet access or valid address for delivery. Delivery of documents is designed for parties who do not have access to the secure internet site, but any user can also be a recipient for document delivery. The stacking order of documents in the package can be easily adjusted, and documents can be easily added or removed for a customized delivery package. Sending a custom delivery package can be a simple matter of clicking and dragging documents into the desired stacking order, selecting secure e-mail, fax, direct upload, or overnight delivery, and clicking send. Like access, the delivery function of the Web-based systems and methods described can be available 24 hours a day, 7 days a week.
- the delivery of documents can be accomplished using fax, SFTP, HTTP, custom interfaces (see step 817), e-mail (see step 818), e-signature (see step 819), electronic transfer delivery (EDT), overnight package delivery via a shipping vendor, or authorized site-to-site delivery (see step 820).
- e-mail such as step 818
- the system will send out an e-mail to the at least one authorized recipients for alerting the recipients to take action of picking up documents from a document pickup portal 821.
- the e-mail may state that the recipient must take action of picking documents within certain time limitation, otherwise the link for picking up documents will expire for security reasons.
- the system can require the input of e-signature within certain time limitation for security reasons.
- the documents will be delivered into the recipient's inbox, where the delivered documents can be processed and indexed depending on the recipient's choice. Of course, there can be more than one recipients regardless means of delivery.
- loan packages can be delivered electronically.
- sending loan packages to external underwriters or investors can be achieved by selecting the loan to send, choosing the documents to be included, and selecting the delivery function of the systems described to send the documents.
- closing packages can be generated automatically or manually based on specific requirements of the intended lenders and/or investors.
- Authorized users 802 and authenticated APIs 804 for accessing the online RlA collaborative intelligent paperless workplace 813 can vary and depend on each particular industry utilizing the systems or methods described.
- An administrator can provide a custom level of security for a user to access particular offices, cabinets, or folders, wherein offices, cabinets, or folders of the systems or methods described can be organized based on different factors such as location, timing, stage of process, etc.
- potential users in the mortgage/loan industry can include due diligence auditors, underwriters, mortgage brokers, secondary marketing, investors, and other post-closing users.
- due diligence processes typically require people to be transported to work where the paper is located. Using the systems and methods described herein, due diligence auditors no longer have to be sent on the road to conduct their review. Analysis can be conducted from any location, making more auditors available to accept assignments without the expense and inconvenience of being on the road.
- Underwriters - using the systems and methods described herein underwriters can handle files more quickly and efficiently, accelerating decisions that once required meetings and phone calls.
- Mortpape Brokers - using the systems and methods described herein account executives and brokers can view the loans' status on line, reducing time-wasting phone calls. Loans can be achieved electronically, saving hundreds of dollars for long-term storage.
- loan can be dragged and dropped from one pool or inventor cabinet to the next, accelerating transaction closing.
- secondary marketing can be notified via e-mail when the loan is ready to be funded, along with other interested parties such as the processor, account executives, or even the broker who send the loan.
- Systems and methods for intelligent paperless document management described herein provide not only a change in mortgage processing, but also a transformation of entire industry. Systems and methods described also can provide additional functions such as workflow enablement, notifications and instant communications of things like loan conditions and status. Systems and methods described can also provided compatibility via Web design for various internet enabled or mobile devices to receive and view necessary information. Any devices generally known to be capable of connecting to the internet and use a Web-based software are within the scope of the invention.
- FIG. 1 depicting the method/apparatus of the instant invention.
- unorganized information is captured by the apparatus from various office devices such a computer, a FAX, an e-mail system, a scanner, or uploaded to a FTP or a Web site 101.
- the captured documents or information, unorganized and unidentified when acquired are organized into an information matrix known as Knowledge Objects by referencing a Knowledge Warehouse 102, and stored in an information data warehouse 103.
- Knowledge Objects are then transformed into Business Objects, such as electronic documents and transaction sets such as MlSMO standard XML files 104.
- the Business Objects arc stored in business data warehouse or delivered to users of the system and external organizations 105.
- a Work-flow and Decision engine uses the Business Objects to facilitate both manual and automated business decisions, and collaboration 106.
- Figure 2 depicting one preferred embodiment of the method/apparatus of the instant invention. It will be recognized by those of skill in the art that Figure 2 is only one example or embodiment of the instant invention; other embodiments of the instant invention may be recognized by reference to Figure 2 and/or the description herein. For example, each of the steps described in Figure 2 may be modified; further, many of the steps are optional so that one or more steps may be eliminated. Also, other steps may be added. Similarly, the order ofthe steps may be changed or rearranged in numerous ways. Each of these embodiments is within the spirit and scope of the invention as defined by the appended claims. Capture documents:
- Pages, documents, sets of documents, a Dox Package, or Dox Packages are sent electronically to the system for classification or/and extraction of data 200.
- Such documents may be input in any sequence and by or through any manner known to those of skill in the art such as from a fax machine, scanner, c- mail system or any other electronic communication device.
- the document or documents may be in text, electronic, paper, or image form, or a mixture of formats. If needed, in preferred embodiments, the documents) are captured by techniques or in a manner known to those of skill in the art.
- the Dox Package is separated into Image type documents or Text type documents as they are captured.
- Image pre-processing :
- the document image quality in particular, from documents obtained by low resolution scans or facsimile transmission may not be good enough for direct OCR. Therefore, primary image processing may optionally be done to bring the image to the requisite quality for OCR, and Image based classification 202.
- noise is removed from the image by technologies such as de-skew techniques and de-speckle techniques, a change or changes in DPI, and/or image registration correction or by a combination of the above and/or similar techniques.
- IBC Image Based Classification
- the IBC 203 attempts to identify one or more discrete pages using Image Based Features like lay out white space distribution, and other features registered in the collection of document feature descriptors by the Document Learner.
- OCR Optical Character Recognition
- the portions of or the entire image of a page or document is converted into text using OCR by means known in the art 204.
- the OCR program is available commercially.
- the OCR engine is supported with a general as well as a business-domain-specific dictionary to increase the accuracy.
- the OCR output may optionally be in text and xml formats, or may be in other formats.
- the output file is converted to a grid- based matrix format to form a text-pictorial representation of the document (Document Grid File).
- Text type documents such Microsoft word documents are also converted into a Document Grid File.
- the output from image preprocessing and OCR is used for the Image Quality Detection IQD 2OS. Identification:
- the method/apparatus attempts to identify one or more reference pages and then the documents are grouped logically based on the reference page and/or affinity 206.
- Numerous classification engines known in the art can be used, separately or together, including a Word Map 207, a SVM classifier 208, a Location Diagram 209, a Bayesian classifier 210, and a critical- feature-based identifier 211, but any manner known to those of skill in the art can be used.
- the engines are used in a particular order.
- the result is accepted as the identification, and taxonomy classification, of the document; in other or the same preferred embodiments, if most of the classification engines agree, then the result is accepted as the identification of the document. If the document, page or sheet is not identified at this point, further analysis is performed with the aid of a human collaborator (i.e., via escalation). Further, in preferred embodiments, discontinuities are used to identify pages from a single document, e.g., a sentence or a table separated by a page break.
- each document page is classified into, or sorted into, one of the taxonomy classes, as defined by the MISMO standards committee or pre-programmed by user using the document learner, or a class designated by a human collaborator. If the putative class identified by the system is unknown and it cannot be classified by the system, the document or representation thereof may be stored in a feedback folder for further manipulation. Taxonomy classification is also done in multiple levels to identify class, sub-class, and version of the document. Taxonomy classification is preferably performed using multiple classification engines. AU the outputs of the taxonomy classifier may be flagged or designated as one of four types: (1) classified, (2) multi-class, (3) ambiguous, (4) unknown. Document pages flagged as unknown are submitted to an OCR program from a different manufacturer 212, and re-identified 213 using the same Identifier Engines, 207, 208, 209, 210, 211.
- Verifier The pages or documents thus far classified may be further evaluated automatically. In preferred embodiments, those documents that fall into categories 2 and 3 are forwarded to the verifier. Verifier:
- the Information Sequencing process is used to create a sequence matrix from the information acquired during the previous steps. 216.
- Document pages that are still not mapped to a taxonomy class due to bad image quality, a new variation of a document, or for other reasons that do not result in immediate identification or classification are flagged as Unknown.
- Document pages that fall below the confidence threshold value that may be preset or varied by the user, even after the verifier, are sent to exception handling client (Classification client) (i.e., via escalation) 218.
- Classification client i.e., via escalation
- human collaborators can verify the class, assign a class, or note that the document cannot be identified. If a human collaborator verifies or changes the class, this information is sent to a feedback box for an incremental learning.
- the human collaborator is presented with an image comprising the header and footer region of the page or document in question, and optionally with an image of the entire page if the image quality is poor.
- the identification of the page or sheet may be made in reference to the header and footer information, although display parameters, such as position of the various images on the human collaborator's computer screen and zoom capabilities, may be varied by the human collaborator.
- Escalation may occur before, but preferably occurs after, the verifier step. Apply filter for classification:
- a filter may optionally be applied.
- An example is if two documents are very close in format and data, but they differ in a very specific property and because of that they belong to different class.
- a weighted filter that is a Location Diagram with primary key set for the distinguishing property or feature, is applied so that those can be classified accurately and rapidly.
- This technique is also used for determination of different versions of documents. For example, two notes may have very similar contents but differ in specific feature such as the absence or presence of an interest rate adjustment clause, need to be put in different classes for business decisions involving an Adjustable Rate Note and/or a Fixed Rate Note. Collating (class specific):
- document pages are collated using methods of analysis based on Location Diagrams and Feature Vectors as may be understood by one of skill in the art 219. These methods of analysis determine the sequence, page numbers of pages and sheets within documents. This process of collation is capable of determining not only the class of a document page (which in most cases is determined earlier at the Classification step or the Verifier step) but also the exact identification of a document including the version number of each particular document within the Dox Package. The collation methods also correctly identifies the pagination within the document, and also notes and records the presence of duplicate documents.
- the method/apparatus of the instant invention may find and note as identical two identical mortgage notes in a single taxonomy class.
- This collation process of the instant invention is differentiated from classification technologies known in the art by its ability to distinguish closely related documents.
- An example of this is that the method/apparatus of the instant invention can pick two mortgage notes out of a Dox Package, correctly paginate them, and identify and log them as separate, but otherwise identical, documents. Pages or documents are then segregated into a logical group determination, and the pages are mapped to a predetermined business-specific or user-determined taxonomy.
- the collation process is based on incremental learning and various artificial intelligence ("AI")-based techniques, which may include one or more of the following, such as:
- the collate confidence matrix which is the result of the above-described collation process is preferably used for final formation of documents.
- the collate confidence matrix represents affinity among various pages, positions of the pages within sets and the confidence of mapping to a particular taxonomy.
- extraction of information or data from the documents or Dox Package that has been captured using the method or apparatus of the invention is first done automatically from readily identifiable fields 221 and image snippets of other fields location are re-submitted to the OCR step with a field specific dictionary before repeating the extraction process 220.
- Using a Location Diagram-based method allows the location of fields even in case of variation between pages or documents within or between Dox Packages 221. Values missed by automatic extraction of these methods may be located by an automatic field locator 222.
- the automatic field locator uses auto field location based on Location Diagrams 223 and Image based field locator 224.
- the region of the page and/or sheet in question is escalated and the field may be identified with the assistance of a human collaborator by escalation 225.
- the human collaborator may be shown only the relevant region of a page or sheet (Image Snippet) and may identify the region containing the data to be extracted by simply mousing over the region with the values extracted by further processing 226 and, in preferred embodiments, the location of the value within the document then sent to the feedback folder for future reference in regard to learning and optimization of the system.
- the human collaborator indicates exactly where the field is located.
- Relevant information is extracted from documents that have been successfully classified.
- an overall weight may be associated with that solution and used to improve future selection of fields in a particular class of document. Scrubbing and verification of extracted information:
- Extracted values are scrubbed to get exact value 227. Scrubbing further transforms the extracted value to a specific data type. The accuracy of the scrubbed value is verified. Thus, the system provides multiple confidence levels for decision-making. The system generates a Knowledge Object from the scrubbed results. The values with very high extraction confidence but very low scrubbing confidence are sent to a human assisted Field Location Process ("Manual FLP") 225. The system generates a field value from scrubbed results that pass the confidence threshold for the overall process. Extraction exception handier:
- Extracted data falling below confidence threshold value is sent to exception handling client (Manual Extraction Process ("MEP")) ('•&, via escalation) 225 & 226.
- MEP Manual Extraction Process
- Human collaborators can verify and/or change the data and/or extracted information in reference to the Dox Package.
- each field subjected to MEP is extracted by a minimum of two human collaborators and the system compares the extracted value. In the event of a discrepancy, the value in question can be sent to additional human collaborators.
- the method/apparatus of the instant invention may optionally keep track of which data was viewed by human collaborators, and how long they viewed the data, in order to detect potential fraud or illicit activities. Information related to exceptions may also be used for statistical learning.
- the human collaborator mouses over the exact value to be extracted. This is referred to herein as a "snippet" or a "text snippet” and die method/apparatus can pull the snippet and subject it to further scrubbing and processing 227. These snippets of required/specific values may also be extracted and used for formation of Knowledge Objects. Transformation:
- the processes preferred embodiments of the instant invention typically extracts the fields (as they appear in the document) required for various business and/or compliance requirements, then transforms them into facts that can be used further for decision making by an automated rules engine or search engine by packaging these facts and other related information such as text and image snippets, x,y coordinate location of these facts from a Location Diagram into an entity referred to herein as a Knowledge Object.
- a Knowledge Object 228 is an information matrix with the relationship among all the information entities clearly defined 229.
- Knowledge Objects created for use in particular business context such as a MISMO SMART Doc XML file, custom transaction set or electronic document.
- Business Objects give data a business centered view of the information captured by the method/apparatus.
- 230 Business Objects are stored and used for business decision making by a Decision Engine.
- These Knowledge Objects and Business Objects are stored in an electronic data repository which can further be used by a decision engine, 231 a rule engine, or a search engine to make various decisions and/or accelerate, support, or validate decisions.
- Figure 4 depicts Business Object formation.
- the relationships among all Knowledge Objects is established by a method called Collation 401.
- the output of collating Knowledge Objects is done by referring to a knowledge map which has a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a knowledge map which has a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- a business-process-specific knowledge representation of the Business Object required for making business decisions 402.
- Figure 4A depicts the relationships among the Knowledge Objects. The relationship between
- Knowledge Object P 1 410 and Knowledge Object P2 411 is shown in the figure.
- Il and 13 is the set of common features belonging to Pl and P2
- 12 is the set of data elements
- 14 is set of location co-ordinates (snippets and regions)
- 15, 116, 117, and 18 are the other attributes of P2. Since 11 and 13 are common to Pl and P2, the knowledge map is referenced to determine if they have affinity to the same category of Business Object such as a Promissory Note.
- Figure 4B depicts the process of Dox Package creation.
- Document-1, Document-2, Document-n 412 413 414 have their individual attributes. (Attributes from left to right 415 416.) Based on these attributes these documents are mapped to the taxonomy.
- collation process is used to determine affinity to a Dox Package based on common attributes such as Loan number or borrower name.
- Figure 4C depicts the process of Document formation. Here pages page-1 to page-n 417 418419 420 421 based on closeness among pages. Feature Vectors and affinity 422 are mapped to different documents and their copies, revisions.
- Figure 4D depicts a Business Object MISMO AUS Transaction set 428.
- the Knowledge Objects extracted from various forms like 1003 429, 1004 434, and Note 433 are combined to form a transaction set for underwriting of a loan using a rules engine.
- Incremental Learning :
- the system of preferred embodiments of the instant invention performs incremental learning and tuning based on feedback and/or unclassified documents.
- AU Feature Vectors are retuned without actual calculation of relative distances.
- the incremental learning is based on statistical analysis of exception and tuning.
- the system keeps watch on statistical data of the collate, classification and extraction to dynamically tune various control parameters and optimize results. Further, in preferred embodiments, the method/apparatus can readily keep track of where human collaborators reviewed data and how long they accessed the data, thus enabling an operator of the instant invention a certain level of protection against fraud. Learning:
- Figure 3 depicts the flow of learning in one embodiment of the instant invention.
- the document samples for the document to be learned, and document-specific dictionaries and generic, as well as domain, dictionaries are loaded in to the Learner's Knowledge Base. 301.
- the Learner reads the document samples, and if document specific dictionaries are not available, then one is generated from the sample documents. 302.
- human input such as very specific key phrases and location are provided for learning, if required.
- Text Feature Vectors are created using image processing, machine learning and Location Diagram based techniques and other methods known in a manner known to those of skill in the art.
- the Feature Vector represents various text features including frequencies., relative locations and Location Diagrams. 304.
- the distances among the Feature Vectors representing different classes, locating different information are maximized. Weights are assigned to Feature Vectors based on their uniqueness and distance from the other Feature Vectors. 305. Using Statistical techniques thresholds are calculated. 306. If the Feature Vectors uniquely identify document 307 document is flagged and Feature Vectors are loaded 308. Otherwise the Feature Vector is re-tuned to prevent misclassification by maximize the distance to from the wrongly classified document class 309. Similarly Feature Vectors are created based on image features. 310. These sets of Feature Vectors are then mapped to a class. 311. The Feature Vectors are tuned to optimize the results. 312. The documents are flagged and corresponding Feature Vectors are loaded in the system. 313.
- the text and image based learning process complements each other and can be performed in any order.
- the output of the learner is a collection of reference-sets that are then stored in a Knowledge Base of the Classifier and Extractor methods to reference.
- The system can prepare reference-sets of known classes with title of the class, i.e., taxonomies.
- the system can use either a dictionary specific to the endeavor domain (i.e., real estate) or a dictionary specific to a document classes.
- the system can, based on reference-sets, generate a dictionary for each class.
- This dictionary also contains a weight for each word.
- the weight for each word plus a weight for combinations of words is determined based on frequency and Bayesian analysis of word features with reference to document identity.
- Learning also generates Feature Vectors based on Location Diagrams for each set (referencc-sct).
- the Feature Vectors generated represent precisely that set of documents, or at least most of the documents, in that reference-set.
- the method/apparatus can maximize distance between Feature Vectors derived from Location
- the method/apparatus can also load Feature Vectors from an outside source.
- the instant invention features in one embodiment a method/apparatus that identifies and collates individual documents and revisions of the same document type within a set.
- the method/apparatus automatically identifies discontinuilies using various algorithms to identify groups or logical units of documents.
- the instant invention takes advantage of its computer and human collaboration and to utilize the strengths of both.
- the output of the method/apparatus is a Business Object like MISMO Smart Docs.
- the Business Object is a business- centered Knowledge Object representation useful to a business decision maker.
- the method/apparatus of preferred embodiments of the instant invention has a method for making decisions based on business processes to select and organize the Business Objects and provide automated decisions in some situations.
- the Business Object contains a complete collated and bucketed set of documents, complete relationships of KOs for specific process, etc. Further collated documents and information is presented with the business identification furnished and mapped to the business-specific or user-provided taxonomy. Further this method/apparatus measures relatedness among various pages and sorts and identifies documents on the principle of the reference page.
- the instant invention in preferred embodiments, collates pages from the input set of documents into a logical/numerical sequence.
- the fields required for different business processes are extracted from these collated and taxonomy-mapped buckets.
- the instant invention also provides for fact transformation so that the information extracted from the pages in the document set is converted into usable form and can be used directly according to various business-specific manners.
- the instant invention in preferred embodiments, provides the formation of Knowledge Objects and additionally ready-to-use Business Objects.
- the processes of the instant invention typically extract the information fields required for various business and/or compliance requirements, then transforms them into facts that can be used further for decision making.
- the decision system used for analyzing the document set provides logical decisions based on the information within and the business context.
- the instant invention offers a collation system and complete organization and fact extraction solution that forms the information matrix, Knowledge Objects. This allows information flow from paper documents from a wide variety of types of images to decision-making based on error free analysis using the techniques of intelligent mapping available to the operator of the instant invention.
- the invention is highly scalable because of its dynamic learning ability based on Feature Vectors and ability to create Business Object based on requirement and business process.
- Dox Packages as Business Objects (e.g. MlSMO SMART Doc) that is business- type specific (Underwriting, Servicing, Closing process etc.);
- one of (he major purposes of the process of the instant invention is Knowledge Object and Business Object creation.
- the final output is a Business Object and not only a set of classified or sorted documents.
- the purpose of the instant invention is not classification or extraction but to create Business Objects like MISMO Smart-Docs from Knowledge Objects, thereby accelerating automated and manual business decisions.
- the flow of preferred embodiments of the instant invention is uniquely valuable in yielding Business Objects.
- Various algorithms are used in a manner and sequence to obtain optimal accuracy.
- the process of preferred embodiments of the instant invention emphasizes feature/knowledge extraction out of Dox Packages with classification and document separation an allied output.
- the instant invention in preferred embodiments, locates the knowledge portion within a Dox Package irrespective of slippage and page numbers.
- the instant invention may provide information for the downstream business process directly from Dox Package capture to Business Object creation and decision-making based on the Business Objects.
- the intelligent information locator of the instant invention may help the business process by locating the business critical information.
- the location algorithm uses a novel method to provide accuracy.
- the method of preferred embodiments of the instant invention identifies all available sources and multiple occurrences of the same information across the Dox Package, i.e., to different versions of the same type document; this enables the user to compare this information and make decisions based on the most recent or relevant information.
- the image- and text-based information locator of the instant invention takes advantage of image and text properties of the documents while locating the information.
- the instant invention recognizes that the document boundaries in business context are not as significant as the multiple occurrences and sets of Knowledge Objects that suggest the presence of more than one form of the same types.
- the information locator may also indicate versions and facilitates relevant decisions.
- the Location Diagram-based method nay be used for rapid location of data, and, which in turn, returns the data association with the image.
- the Location Diagram based method locates may collect information from proper page irrespective of similarities among the pages, as well as new variations among the forms.
- the Location Diagram-based locator can locate appropriate information based on the version of the form.
- the Image and Location Diagram based locator can locate the information on forms irrespective of poor quality of images/OCR output.
- the system of the instant invention either may extract or make available the relevant portion or the Dox Package for knowledge extraction by an operator by increasing extraction efficiency by up to 5X over prior methods.
- the instant invention in preferred embodiments, features less turnaround or learning time.
- the instant invention in preferred embodiments, features incremental learning as to locations.
- the instant invention in preferred embodiments, features automatic and semiautomatic learning for added flexibility.
- the instant invention in preferred embodiments, features the verifier for verifying location.
- the instant invention in preferred embodiments, features a scrubber which can scrub extraction output.
- the instant invention features the ability of establishing knowledge- based relationship among all the relevant knowledge portions resulting in a rich Knowledge Object that can help in Creation of Business Objects.
- the instant invention in preferred embodiments, features collation of Knowledge Objects to create Business Objects.
- the instant invention in preferred embodiments, features efficient decision making based on Business Objects.
- the instant invention features a method of doing business by processing a Dox Packages wherein each Dox Package has at least two pages wherein minimal human intervention is involved in the extraction of information and/or data.
- the Dox Package lias documents related to a mortgage.
- the instant invention features a method of doing business by processing a group of Dox Packages wherein each Dox Package has at least two pages wherein the information is extracted from the Dox Packages and organized ten times as fast as a human operator.
- the Dox Package has documents related to a mortgage.
- the instant invention features a method of doing business by processing a group of documents, i.e., a Dox Package, where the process comprises some or all of the following steps:
- the instant invention features a method of doing business by processing a group of documents using a computer where the process comprises some or all of the following steps:
- the instant invention features an apparatus for analyzing a group of documents using the methods described herein wherein said apparatus comprises a computer.
- the instant invention features an apparatus for processing a group of documents where the apparatus performs all or some of the following steps:
- the instant invention features a method of analyzing a bundle of loans assembled for sale on the secondary market wherein over 30%, over 40%, over 50%, over 60%, or over 70% of the mortgage documents are analyzed and the data/information is extracted.
- ambiguities in the processing of the documents are escalated to a human collaborator, in particular this may occur during or following the classification step, the field location step, and/or the data extraction step.
- the step of performing optical character recognition is performed by, or with the assistance of, a computer.
- the step of classifying is performed by, or with the assistance of, a computer.
- the step of verifying is performed by, or with the assistance of, a computer.
- the step of collating is performed by, or with the assistance of, a computer.
- the step of extracting data is performed by, or with the assistance of, a computer.
- the step of scrubbing is performed by, or with the assistance of, a computer.
- the outputting is performed by, or with the assistance of, a computer.
- the step of displaying is performed by, or with the assistance of, a computer.
- ambiguities at any step are escalated to a human operator.
- the group of documents being analyzed is a group of mortgage loan documents.
- the groups of documents being analyzed may be home appraisals, credit reports, and a single loan file where it is frequently used for underwriting purposes.
- the invention also features a method of operating a business where a purpose of the business is to offer the method/apparatus of preferred embodiments of the instant invention as a service.
- the instant invention features advertising the method/apparatus of the instant invention and/or advertising the availability of a service featuring the method/apparatus of the instant invention.
- Figure 5 depicts screen shots of output obtained through the use of one embodiment of the instant invention. That is, using a Dox Package analyzed by the method/apparatus as described herein, the following exemplary data was available for analysis for making business decisions.
- the brokers do not have to change their behavior in order to deliver the native pdf package from their desktop loan origination system (LOS). Instead of printing to their HP printer, they select a special print driver and when they click "print," the information flows electronically and securely into the system described.
- LOS desktop loan origination system
- a new user can go to the katalyst.net site for setting up a new Client Site- for a new company ( Figure 9).
- Each company can have a unique name. If a particular name has been used, the program will ask the user to input a different name.
- Figure 10 Figure 11 shows that a new company name has been accepted by the program. After entering into the program, as the first step of process, the program will ask the user to provide a unique site address in order to login in die future (Figure 12).
- Figure 13 shows that the user can input info for the company as the second step of the process.
- Figure 14 shows that the user can input billing info as the third step of the process.
- An administrator can be designated as the fourth step of the process ( Figure 15).
- the desktop tool Messenger has three major functions - scan, print, and pdf upload files for uploading documents into the Web-based system (Figure 20).
- Figure 21 shows the scan function of Messenger.
- Figure 22 shows the print function of Messenger.
- Figure 23 shows the pdf upload function of the Messenger.
- Messenger allows a folder to be created or designated for uploading documents via scan, print, or pdf upload (Figure 24).
- Figure 25 shows that a demo document is being uploaded and bypassing auto-indexing. The user can monitor the progress of uploading the document as shown in Figure 26.
- Figure 27 shows that the demo document has been successfully uploaded.
Abstract
Description
Claims
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/US2007/086673 WO2009073032A1 (en) | 2007-12-06 | 2007-12-06 | Systems and methods for intelligent paperless document management |
Publications (2)
Publication Number | Publication Date |
---|---|
EP2227750A1 true EP2227750A1 (en) | 2010-09-15 |
EP2227750A4 EP2227750A4 (en) | 2012-06-20 |
Family
ID=40718013
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP07865320A Ceased EP2227750A4 (en) | 2007-12-06 | 2007-12-06 | Systems and methods for intelligent paperless document management |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP2227750A4 (en) |
CA (2) | CA2957327A1 (en) |
WO (1) | WO2009073032A1 (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130332403A1 (en) * | 2012-06-12 | 2013-12-12 | International Business Machines Corporation | Leveraging analytics to propose context sensitive workflows for case management solutions |
WO2015008300A2 (en) * | 2013-07-19 | 2015-01-22 | Parag Kulkarni | A system for instance specific, device-specific, duration-specific, view-specific, time-stamp-specific, and network specific, file/content sharing |
US20180232708A1 (en) * | 2014-11-04 | 2018-08-16 | Beacon Communications, Llc | Consumer service portal |
EP3018592B1 (en) | 2014-11-04 | 2018-03-21 | Tata Consultancy Services Limited | A computer implemented system and method for managing a stack containing a plurality of documents |
WO2019036040A1 (en) * | 2017-08-18 | 2019-02-21 | ISMS Solutions, LLC | Computer based learning system for analyzing agreements |
US11049042B2 (en) * | 2018-11-05 | 2021-06-29 | Convr Inc. | Systems and methods for extracting specific data from documents using machine learning |
US11270213B2 (en) | 2018-11-05 | 2022-03-08 | Convr Inc. | Systems and methods for extracting specific data from documents using machine learning |
US10657603B1 (en) * | 2019-04-03 | 2020-05-19 | Progressive Casualty Insurance Company | Intelligent routing control |
CN110909196B (en) * | 2019-10-28 | 2022-07-01 | 北京光年无限科技有限公司 | Processing method and device for identifying inner page cover switching in picture book reading process |
US20230105207A1 (en) * | 2021-10-06 | 2023-04-06 | Bank Of America Corporation | System and methods for intelligent entity-wide data protection |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030217034A1 (en) * | 2002-05-14 | 2003-11-20 | Shutt Michael J. | Document management system and method |
US20050226261A1 (en) * | 2004-04-13 | 2005-10-13 | Satyam Computer Services Ltd. | System and method for automatic indexing and archiving of paper documents |
US20060059162A1 (en) * | 2000-12-27 | 2006-03-16 | Rizk Thomas A | Method and system to convert paper documents to electronic documents and manage the electronic documents |
US7197703B1 (en) * | 2001-04-09 | 2007-03-27 | Critical Technologies, Inc. | System and methodology for the storage and manipulation of documents |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6738740B1 (en) * | 2000-05-31 | 2004-05-18 | Kenneth Barash | Speech recognition system for interactively gathering and storing verbal information to generate documents |
US6904416B2 (en) * | 2001-03-27 | 2005-06-07 | Nicholas N. Nassiri | Signature verification using a third party authenticator via a paperless electronic document platform |
NO316480B1 (en) * | 2001-11-15 | 2004-01-26 | Forinnova As | Method and system for textual examination and discovery |
MXPA06004104A (en) * | 2003-10-09 | 2007-03-21 | Corelogic Systems Inc | Automated financial transaction due diligence systems and methods. |
US20050209955A1 (en) * | 2004-03-16 | 2005-09-22 | Underwood Timothy J | Apparatus and method for document processing |
-
2007
- 2007-12-06 CA CA2957327A patent/CA2957327A1/en not_active Abandoned
- 2007-12-06 WO PCT/US2007/086673 patent/WO2009073032A1/en active Application Filing
- 2007-12-06 EP EP07865320A patent/EP2227750A4/en not_active Ceased
- 2007-12-06 CA CA2745712A patent/CA2745712C/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060059162A1 (en) * | 2000-12-27 | 2006-03-16 | Rizk Thomas A | Method and system to convert paper documents to electronic documents and manage the electronic documents |
US7197703B1 (en) * | 2001-04-09 | 2007-03-27 | Critical Technologies, Inc. | System and methodology for the storage and manipulation of documents |
US20030217034A1 (en) * | 2002-05-14 | 2003-11-20 | Shutt Michael J. | Document management system and method |
US20050226261A1 (en) * | 2004-04-13 | 2005-10-13 | Satyam Computer Services Ltd. | System and method for automatic indexing and archiving of paper documents |
Non-Patent Citations (1)
Title |
---|
See also references of WO2009073032A1 * |
Also Published As
Publication number | Publication date |
---|---|
CA2745712A1 (en) | 2009-06-11 |
CA2745712C (en) | 2017-03-21 |
CA2957327A1 (en) | 2009-06-11 |
WO2009073032A1 (en) | 2009-06-11 |
EP2227750A4 (en) | 2012-06-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8176004B2 (en) | Systems and methods for intelligent paperless document management | |
CA2745712C (en) | Systems and methods for intelligent paperless document management | |
US7747495B2 (en) | Business method using the automated processing of paper and unstructured electronic documents | |
US20190286898A1 (en) | System and method for data extraction and searching | |
US8571974B1 (en) | Methods and systems for automated data collection and analysis | |
US10096064B2 (en) | Method and system for source document data entry and form association | |
US7895166B2 (en) | Automatic document exchange with archiving capability | |
US8180713B1 (en) | System and method for searching and identifying potential financial risks disclosed within a document | |
US8706614B2 (en) | Systems and methods for automated political risk management | |
US9507758B2 (en) | Collaborative matter management and analysis | |
JP6307745B2 (en) | Accounting system | |
US20210209551A1 (en) | System and method for automatic analysis and management of a workers' compensation claim | |
Miller | All text considered: A perspective on mass digitizing and archival processing | |
Saund | Scientific challenges underlying production document processing | |
US20040117404A1 (en) | System for utilizing audible, visual and textual data with alternative combinable multimedia forms of presenting information for real-time interactive use by multiple users in differnet remote environments | |
US20050210068A1 (en) | Title examination systems and methods | |
CA2914591C (en) | Systems and methods for intelligent paperless document management | |
US20090132406A1 (en) | System and method for paperless loan applications | |
Schuler | E-discovery: creating and managing an enterprisewide program: a technical guide to digital investigation and litigation support | |
JP4024267B2 (en) | Supplier guidelines system | |
US20220138259A1 (en) | Automated document intake system | |
US20050209872A1 (en) | Title quality scoring systems and methods | |
Derrig et al. | Effective Document Review Techniques in Eclipse and Relativity | |
Fernando | Intelligent Document Processing: A Guide For Building RPA Solutions | |
US20220335087A1 (en) | Data processing apparatus, data processing system, and data processing method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 20100617 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC MT NL PL PT RO SE SI SK TR |
|
AX | Request for extension of the european patent |
Extension state: AL BA HR MK RS |
|
RIN1 | Information on inventor provided before grant (corrected) |
Inventor name: MALANEY, SANJEEV Inventor name: VISWANATHAN, KRISHNWAMI Inventor name: MALANEY, VIKRAM Inventor name: EVANS, JEFFREY Inventor name: KULKARNI, PARAG |
|
DAX | Request for extension of the european patent (deleted) | ||
A4 | Supplementary search report drawn up and despatched |
Effective date: 20120523 |
|
RIC1 | Information provided on ipc code assigned before grant |
Ipc: G06F 17/30 20060101ALI20120516BHEP Ipc: G06F 17/00 20060101AFI20120516BHEP |
|
17Q | First examination report despatched |
Effective date: 20161007 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R003 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED |
|
18R | Application refused |
Effective date: 20180420 |