System and method for automated processing of electronic documents

• US 10,339,376 B2
• Filed: 09/09/2014
• Issued: 07/02/2019
• Est. Priority Date: 06/25/2014
• Status: Active Grant

• Current Assignee: Cognizant Technology Solutions India Private Limited (Cognizant Technology Solutions Corp.)
• Original Assignee: Cognizant Technology Solutions India Private Limited (Cognizant Technology Solutions Corp.)
• Inventors: Pati, Peeta Basa, Nagamalla, Srinivas, Kongara, Harish, Rao, Ejiribidi Rajeshekhar, Veera, Ravi
• Primary Examiner(s): Hong, Stephen S
• Assistant Examiner(s): Anderson, Broderick C
• Application Number:
  US14/481,296
• Publication Number:
  US 20150379342A1
• Time in Patent Office: 1,757 Days
• Field of Search: 715243
• CPC Class Codes: G06V 10/225 based on a marking or identifier characterising the area
• CPC Class Codes: G06V 30/414 Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text
• CPC Class Codes: G07D 7/12 Visible light, infrared or ultraviolet radiation

Abstract

In accordance with an aspect of the present invention, a system and method for automated processing of electronic documents is provided. The said system comprising a precursor module configured to receive an electronic document and convert into a textual equivalent; a data ascertainment module configured to identify the textual snippets, more particularly, the demarcated sections, corresponding to the at least one structured section and logically separate the demarcated sections the remaining textual equivalent; and, a pass-receiving module configured to receive said logically separated demarcated sections and the remaining textual equivalent and capture business relevant data committed therein, wherein said pass-receiving module captures the business relevant data from the demarcated sections by locating at least one master trigger and at least one proximate co-trigger in the demarcated sections and geometrically coupling said located triggers and co-triggers into at least one potential zone and extracting the business relevant data committed within these potential zones.

Claims

1. A system for automatically processing electronic documents, the system comprises:

a memory comprising programming instructions;

a processor configured to execute the programming instructions stored in the memory and configured to;

receive an electronic document comprising at least one of;

a structured section or an unstructured section;

convert the electronic document into a textual equivalent;

scan the textual equivalent and demarcate those sections that correspond to one or more predetermined structural attributes;

separate the one or more demarcated sections from the textual equivalent and retrieve the one or more demarcated sections corresponding to the structured sections and a remaining textual equivalent corresponding to the unstructured sections as distinct inputs;

receive the one or more demarcated sections and the remaining textual equivalent as the distinct inputs;

identify one or more master triggers within the received distinct inputs;

generate one or more potential zones with the identified one or more master triggers, wherein the generated one or more potential zones is defined by at least one geometric shape formed by geometrically coupling the master triggers and co-triggers proximate to the master triggers into the geometric shape such that the master triggers and the co-triggers form one or more vertices of the geometric shape;

generate one or more rules of extraction to determine at least one extraction type from a plurality of extraction types, wherein each of the plurality of extraction types represent a particular method of extraction, based on the type of electronic document, wherein the type of electronic document is ascertainable based on identification of a template type of the electronic document associated with the demarcated section; and

capture the business relevant data contained in the generated one or more potential zones within the one or more demarcated sections and the remaining textual equivalent based on co-ordinates of the vertices of the geometric shape formed by the one or more master triggers and the co-triggers by applying the determined at least one extraction type.

2. The system of claim 1, wherein the processor is configured to receive the electronic document, and convert the electronic document into a XML, document, further wherein the XML Document comprises of one or more XML snippets corresponding to the at least of:

the structured section or the unstructured section.

3. The system of claim 1, wherein the processor is configured to:

scan the textual equivalent for identification of those textual snippets from within the textual equivalent that correspond to the one or more structured sections, wherein the one or more predetermined structured attributes comprise attributes including type of template and name of owner; and

generate the one or more demarcated sections based on-boundaries sketched around the identified one or textual snippets.

4. The system of claim 1, wherein the processor is configured to separate the one or more demarcated sections from the textual equivalent and retrieves the one or more demarcated sections corresponding to structured sections and a remaining textual equivalent corresponding to unstructured sections as distinct inputs by logically separating the demarcated sections from the textual equivalent using a partitioning element.

5. The system of claim 1, wherein the processor is configured to:

identify the one or more co-triggers associated with the one or more master triggers, wherein-the master triggers are at least one of;

a word or a phrase that has a high probability of appearing within the one or more demarcated sections, and the co-triggers are at least one of;

a word or a phrase that has a high probability of appearing near the master triggers;

record the identified one or more master triggers that are identified in the one or more demarcated sections and mark occurrences and locations of the recorded one or more master triggers;

retrieve the one or more co-triggers that are likely to co-exist with the recorded one or more master triggers;

determine the one or more proximate co-triggers from the retrieved co-triggers, the proximate co-triggers are those co-triggers that appear near to each of the one or more recorded master triggers; and

couple, geometrically, the one or more identified master triggers with the one or more proximate co-triggers identified to mark the one or more potential zones.

6. The system of claim 5, wherein the identified master triggers are ranked based on one or more predetermined criteria;

and conflicts that arise due to detection of the one or more ranked master triggers at multiple locations of the geometric shape are resolved by perusing through the ranks of the one or more ranked master triggers and-retaining only those ranked master triggers which have better ranks.

7. The system of claim 5, wherein the processor is configured to:

verify presence of the one or more recorded master triggers in the one or more prelisted master; and

verify presence of the one or more proximate co-triggers in the one or more pre-listed co-triggers.

8. The system of claim 1, wherein the processor is configured to capture the business relevant data contained in the generated one or more potential zones within the one or more demarcated sections, and the remaining textual equivalent by applying the at least one determined extraction type comprising:

a co-ordinate based extraction element, wherein the co-ordinate based extraction element is configured to;

generate one or more extraction rules for the one or more potential zones within the demarcated sections and the remaining textual equivalent based on determination of position details including (x, y) co-ordinates of the one or more vertices of the defined geometric region; and

apply the generated one or more extraction rules to the one or more potential zones within the demarcated sections and the remaining textual equivalent, and extract the business relevant data contained therein;

or

a context based extraction element, wherein the context based extraction element is configured to;

generate one or more regular expression patterns for the one or more potential zones identified by one or more vertices of the defined geometric shape within the demarcated sections and the remaining textual equivalent, wherein the one or more regular expression patterns define a set of conditions of extraction which are applied between a pre gazetteer and a post gazetteer such that any data which lies between the pre gazetteer and the post gazetteer, and meets the conditions of the regular expression pattern is identified as business relevant data; and

apply the generated one or more regular expression patterns to the one or more potential zones within the demarcated sections and the remaining textual equivalent, and extract the business relevant data contained therein.

9. The system of claim 8, wherein the processor is configured to:

receive the captured business relevant data; and

convey the extracted business relevant data in accordance with the one or more predetermined instructions, wherein the predetermined instructions are received from one or more users.

10. The system of claim 9, wherein said system is configured to interface with a front-end framework that includes an interactive graphical user interface for facilitating the one or more users to perform one or more operations.

11. A computer implemented method for automatically processing electronic documents, the computer implemented method comprising:

receiving an electronic document comprising at least one of;

a structured section or unstructured section, and converting the electronic document into a textual equivalent;

scanning the textual equivalent and demarcating those sections that correspond to one or more predetermined structural attributes;

separating the one or more demarcated sections from the textual equivalent and retrieving the one or more demarcated sections corresponding to the structured sections and a remaining textual equivalent corresponding to the unstructured sections as distinct inputs;

identifying one or more master triggers within the received distinct inputs;

generating one or more potential zones within the received distinct inputs, with the identified one or more master triggers, wherein the generated one or more potential zones is defined by at least one geometric shape formed by geometrically coupling the master triggers and co-triggers proximate to the master triggers into the geometric shape such that the master triggers and the co-triggers form one or more vertices of the geometric shape;

generating one or more rules of extraction to determine at least one extraction type from a plurality of extraction types, wherein each of the plurality of extraction types represent a particular method of extraction, based on the type of electronic document, wherein the type of electronic document is ascertainable based on identification of a template type of the electronic document associated with the demarcated section; and

capturing the business relevant data contained in the generated one or more potential zones within the one or more demarcated sections and the remaining textual equivalent based on co-ordinates of the vertices of the geometric shape defined by the one or more master triggers and the co-triggers by applying the determined at least one extraction type.

12. The method of claim 11, wherein automatically processing electronic documents comprises, receiving the electronic document and converting the electronic document into a XML document, further wherein the XML Document comprises of one or more XML, snippets corresponding to the at least of the structured sections or the unstructured sections.

13. The method of claim 11, wherein the step of scanning the textual equivalent and demarcating those sections that correspond to one or more predetermined structural attributes comprises:

searching the structural attributes in the textual equivalent, wherein the predetermined structural attributes are used in identifying those textual snippets from within the textual equivalent that correspond to the one or more structured sections, further wherein the one or more predetermined structured attributes comprise attributes including template of a check and form type of an invoice; and

generating one or more demarcated sections by sketching boundaries around the identified one or more textual snippets.

14. The method of claim 11, wherein the step of generating one or more potential zones within the one or more demarcated sections comprises:

developing a repository to include the one or more master triggers and the one or more co-triggers associated with the one or more master triggers, wherein

the one or more master trigger is one of;

a word and a phrase that has a high probability of appearing within the one or more structured sections, and

the co-trigger is one of;

a word and a phrase that has a high probability of appearing near the master triggers;

recording the one or more master triggers that are identified in the one or more demarcated sections and marking occurrences and locations of the recorded one or more master triggers;

retrieving the one or more co-triggers, wherein the retrieved one or more co-triggers are those co-triggers that are likely to co-exist with the recorded one or more master triggers;

determining the proximate co-triggers from the retrieved co-triggers, wherein the proximate co-triggers are those co-triggers that appear near to each of the one or more recorded master triggers; and

coupling, geometrically, the one or more identified master triggers with the one or more proximate co-triggers at the one or more vertices of the geometric shape, to mark one or more potential zones.

15. The method of claim 11, wherein the step of capturing the business relevant data contained in the generated one or more potential zones within the one or more demarcated sections and the remaining textual equivalent by a co-ordinate based extraction technique comprises:

generating one or more extraction rules for the one or more potential zones within the demarcated sections and the remaining textual equivalent based on determination of positional details including (x, y) co-ordinates of the one or more vertices of the defined geometric shape; and

applying the generated one or more extraction rules to the one or more potential zones within the demarcated sections and the remaining textual equivalent, and extracting the business relevant data contained therein.

16. The method of claim 11, wherein the step of capturing the business relevant data contained in the one or more potential zones identified by the one or more vertices of the defined geometric shape within the one or more demarcated sections and the remaining textual equivalent by a context based extraction technique comprises:

generating one or more regular expression patterns for the one or more potential zones within the demarcated sections and the remaining textual equivalent, wherein the one or more regular expression patterns define a set of conditions of extraction which are applied between a pre gazetteer and a post gazetteer such that any data which lies between the pre gazetteer and the post gazetteer, and meets the conditions of the regular expression pattern is identified as business relevant data; and

applying the generated one or more regular expression patterns to the one or more potential zones within the demarcated sections and the remaining textual equivalent, and extracting the business relevant data contained therein.

17. The method of claim 11, wherein the step of capturing the business relevant data includes delivering the extracted business relevant data, further wherein the delivering comprises conveying the extracted business relevant data in accordance with one or more predetermined instructions, wherein the predetermined instructions are received from one or more users.

18. A computer program product comprising a non-transitory computer readable medium having a computer readable program code stored thereon, the computer-readable program code comprising instructions that, when executed by a processor, cause the processor to:

receive an electronic document comprising at least one of;

a structured section or an unstructured section,

convert the electronic document into a textual equivalent;

scan the textual equivalent and demarcate those sections that correspond to one or more predetermined structural attributes; and

separate the one or more demarcated sections from the textual equivalent and retrieve the one or more demarcated sections corresponding to the structured sections and a remaining textual equivalent corresponding to the unstructured sections as distinct inputs;

receive the one or more demarcated sections and the remaining textual equivalent as the distinct inputs;

identify one or more master triggers within the received distinct inputs;

generate one or more potential zones within the received distinct inputs with the identified one or more master triggers, wherein the generated one or more potential zones is defined by at least one geometric shape formed by geometrically coupling the master triggers and co-triggers proximate to the master triggers into the geometric shape such that the master triggers and the co-triggers form one or more vertices of the geometric shape;

generate one or more rules of extraction to determine at least one extraction type from a plurality of extraction types, wherein each of the plurality of extraction types represent a particular method of extraction, based on the type of electronic document, wherein the type of electronic document is ascertainable based on identification of a template type of the electronic document associated with the demarcated section; and

capture the business relevant data contained in the generated one or more potential zones and the remaining textual equivalent based on co-ordinates of the vertices of the geometric shape defined by the one or more master triggers and the co-triggers by applying the determined at least one extraction type.