Intelligent data extraction

• US 10,402,163 B2
• Filed: 02/14/2017
• Issued: 09/03/2019
• Est. Priority Date: 02/14/2017
• Status: Active Grant

• Current Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Original Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Inventors: Ghatage, Prakash, Sampat, Nirav, Viswanathan, Kumar, Thangaraj, Naveen, Meenakshisundaram, Karthik
• Primary Examiner(s): Arjomandi, Noosha
• Application Number:
  US15/432,039
• Publication Number:
  US 20180232204A1
• Time in Patent Office: 931 Days
• CPC Class Codes: G06F 16/31 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 7/02 Comparing digital values G06F7/06, G06F7/22, G06F7/38 take precedence

Abstract

Electronically received data is validated based on a digital data image that is scanned from a paper document. Known paper document source entities, paper document types and associated paper document configuration information are stored in a database. The paper documents are converted to digital data images and optically processed to identify respective source entity and document type information represented within the digital data images. Appropriate document configuration information is retrieved based on association with the detected type of document. Validation target data is extracted from the digital data images based on the configuration information and processed. The electronically received data is validated based on the extracted and processed validation target data.

Claims

1. A method for validating electronic data, the method comprising:

storing in a database, data representative of;

a plurality of entities;

one or more document types, wherein each of the one or more document types is associated one or more of the plurality of entities; and

one or more configurations, wherein each of the one or more configurations comprises configuration information associated with one or more of the document types;

receiving an electronic data validation request, the request comprising electronic data input by a user comprising customer specific data to be validated by comparison with data extracted from paper documents provided by the customer;

receiving digital data images representing electronically scanned paper documents, associated with the electronic data validation request, said paper documents being provided by the customer;

optically processing each respective digital data image to convert the image into machine readable text, identifying from the machine readable text an entity corresponding to one of the stored plurality of entities that is represented within the respective digital data image, and determining a document type corresponding to one of the one or more document types from the respective digital data image;

retrieving configuration information corresponding to each respective digital data image having an identified entity and determined document type based on the identified entity and the determined document type of the respective digital data image;

extracting target data from each respective digital data image for which configuration information is retrieved, wherein the extracting is based on the configuration information corresponding to each respective digital data image;

validating the electronic data input by the user by comparing it with the target data extracted from each respective digital data image of the one or more scanned paper documents associated with the electronic data validation request and determining whether the electronic data and target data match; and

transmitting a response to the electronic data validation request indicating the result of the electronic data validation.

2. The method for validating electronic data of claim 1, wherein the entity represented within the digital data image indicates a source of a corresponding paper document.

3. The method for validating electronic data of claim 1, wherein the configuration information corresponding to an entity and document type indicate types of data, data or data field locations and data formatting information for a digital data image corresponding to the entity and document type.

4. The method for validating electronic data of claim 1, wherein the configuration information is stored in the database as Javascript object notation (JSON) and indicates from which fields of a corresponding respective digital data image to extract the validation target data.

5. The method for validating electronic data of claim 1, further comprising storing the extracted target data in JSON format.

6. The method for validating electronic data of claim 1, wherein the entity of the plurality of entities is identified utilizing natural language processing and data extraction techniques.

7. The method for validating electronic data of claim 1, wherein the document type of the one or more document types is detected from the respective digital data image utilizing fuzzy matching techniques or approximate string matching techniques.

8. The method for validating electronic data of claim 1, wherein the target data is searched and extracted from each respective digital data image for which configuration information is retrieved based on one or more of artificial intelligent techniques, behavior tree searching, fuzzy matching, AhoCorasick algorithm and pattern recognition techniques.

9. The method for validating electronic data of claim 1, further comprising:

creating a new self-learned configuration in instances when the configuration information is not available;

searching within a digital image data document corresponding to the detected document type for numerical expressions that represent various known types of target data based on known target data formatting conventions;

searching for column headers that correspond to the numerical expressions;

searching the column headers and the numerical expressions for keywords and symbols that characterize the data of each column;

storing in the database, the new self-learned configuration, locations and formatting of found types of target data, the column headers, and the data characteristics; and

associating the new self-learned configuration with a detected document type.

10. A system for validating electronic data, the system comprising:

a computer processor;

a memory that stores instructions, wherein the instructions when executed by the computer processor cause the computer processor to;

store in a database;

data representative of;

a plurality of entities;

one or more document types, wherein each of the one or more document types is associated one or more of the plurality of entities; and

one or more configurations, wherein each of the one or more configurations comprises configuration information associated with one or more of the document types;

receive an electronic data validation request, the request comprising electronic data input by a user comprising customer specific data to be validated by comparison with data extracted from paper documents provided by the customer;

receive digital data images representing electronically scanned paper documents associated with the electronic data validation request, said paper documents being provided by the customer;

optically process each respective digital data image to convert the image into machine readable text, identify from the machine readable text an entity corresponding to one of the stored plurality of entities that is represented within the respective digital data image, and determine a document type corresponding to one of the one or more document types from the respective digital data image;

retrieve configuration information corresponding to each respective digital data image having an identified entity and determined document type based on the identified entity and the determined document type of the respective digital data image;

extract target data from each respective digital data image for which configuration information is retrieved, wherein the extraction is based on the configuration information corresponding to each respective digital data image; and

validate the electronic data input by the user by comparing it with the target data extracted from each respective digital data image of the one or more scanned paper documents associated with the electronic data validation request and determining whether the electronic data and target data match; and

transmit a response to the electronic data validation request indicating the result of the electronic data validation.

11. The system for validating electronic data of claim 10, wherein the entity represented within the digital data image indicates a source of a corresponding paper document.

12. The system for validating electronic data of claim 10, wherein the configuration information corresponding to an entity and document type indicate types of data, data or data field locations and data formatting information for a digital data image corresponding to the entity and document type.

13. The system for validating electronic data of claim 10, wherein the configuration information is stored in the database as Javascript object notation (JSON) and indicates from which fields of a corresponding respective digital data image to extract the validation target data.

14. The system for validating electronic data of claim 10, wherein the instructions when executed by the computer processor further cause the computer processor to store the extracted target data in JSON format.

15. The system for validating electronic data of claim 10, wherein the entity of the plurality of entities is identified utilizing natural language processing and data extraction techniques.

16. The system for validating electronic data of claim 10, wherein the document type of the one or more document types is detected from the respective digital data image utilizing fuzzy matching techniques or approximate string matching techniques.

17. The system for validating electronic data of claim 10, wherein the target data is searched and extracted from each respective digital data image for which configuration information is retrieved based on one or more of artificial intelligent techniques, behavior tree searching, fuzzy matching, AhoCorasick algorithm and pattern recognition techniques.

18. The system for validating electronic data of claim 10, wherein the instructions when executed by the computer processor further cause the computer processor to:

create a new self-learned configuration in instances when the configuration information is not available;

search within a digital image data document corresponding to the detected document type for numerical expressions that represent various known types of target data based on known target data formatting conventions;

search for column headers that correspond to the numerical expressions;

search the column headers and the numerical expressions for keywords and symbols that characterize the data of each column;

store in the database, the new self-learned configuration, locations and formatting of found types of target data, the column headers, and the data characteristics; and

associate the new self-learned configuration with a detected document type.

19. A method for validating electronic data, the method comprising:

polling a computer memory queue by robotic process automation for electronic data received via a web-based electronic user filled form for a credit application, wherein the electronic data includes income information;

downloading concurrently a digital data image from an online optical storage when the electronic data is received, wherein the digital data image is based on documents provided by the user and corresponds to the received electronic data;

transmitting the electronic data and the digital data image to a remote optical character recognition system to extract target data from the digital data image and verify that the electronic data matches information that is based on the extracted target data;

extracting target data from each respective digital data image based on retrieved configuration information, wherein the extracting is based on the configuration information corresponding to each respective digital data image and the target data includes income information;

verifying the electronic data by comparing it with the target data extracted from each respective digital data image and determining whether the income information in the electronic data and target data match; and

receiving and displaying the income verification results in a service administrator graphical user interface within a specified time period for further processing of the web-based electronic user filled form.

20. A system for validating electronic data, the system comprising:

a computer processor;

a memory that stores instructions, wherein the instructions when executed by the computer processor cause the computer processor to;

poll a queue by robotic process automation for received electronic data in a web-based electronic user filled form for a credit application, wherein the electronic data includes income information;

download concurrently a digital data image from an online optical storage when the electronic data is received, wherein the digital data image is based on documents provided by the user and corresponds to the received electronic data;

transmit the electronic data and the digital data image to a remote optical character recognition system to extract target data from the digital data image and verify that the electronic data matches information that is based on the extracted target data;

extract target data from each respective digital data image based on retrieved configuration information, wherein the extracting is based on the configuration information corresponding to each respective digital data image and the target data includes income information;

verify the electronic data by comparing it with the target data extracted from each respective digital data image and determining whether the income information in the electronic data and the target data match; and

receive and display the income verification results in a service administrator graphical user interface within a specified time period for further processing of the web based electronic user filled form.