# US 2014/0195891 A1 · System and method for automatically extracting multi-format data from documents and converting into xml

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2014/0195891 A1
- Document ID: US20140195891A1
- Status: Published application
- Kind: A1
- Application number: 13/799,555
- Filed: 2013-03-13
- Published: 2014-07-10
- Earliest claimed: 2013-01-04
- Assignees: Cognizant Technology Solutions India Pvt. Ltd.
- Applicants: Cognizant Technology Solutions India Pvt. Ltd.
- Inventors: Geelapaturu Subrahmanya Venkata Radha Krishna Rao; Parthasarathi Jinka; Venkatesh Srinivasan; Ramesh Selvaraj; Suresh Kumar Ramaswamy; Dinesh Maroo
- Web page: https://patentgrep.com/patent/US20140195891A1

## CPC

- G06F 40/117: Tagging; Marking up (details of markup languages G06F40/143); Designating a block; Setting of attributes (style sheets, e.g. eXtensible Stylesheet Language Transformation [XSLT], G06F40/154)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system, a computer-implemented method and a computer program product for extracting insurance data from one or more documents having one or more file formats and converting into Extensible Markup Language (XML) format is provided. The system comprises a user interface configured to facilitate one or more users to submit one or more documents related to insurance. The system further comprises a business type classification module configured to identify the one or more submitted documents based on a business type. Further, the system comprises a format classification module configured to identify file format of the one or more submitted documents. Furthermore, the system comprises an extraction and conversion module configured to match one or more headers in the one or more submitted documents with one or more pre-stored headers, extract insurance data corresponding to the one or more matched headers and convert the extracted insurance data into XML format.

## Claims

### Claim 1

A system for extracting insurance data from one or more documents having one or more file formats and converting into Extensible Markup Language (XML) format, the system comprising:
a user interface configured to facilitate one or more users to submit one or more documents related to insurance;
a business type classification module configured to identify the one or more submitted documents based on a business type;
a format classification module configured to identify file format of the one or more submitted documents; and
an extraction and conversion module configured to:
match one or more headers in the one or more submitted documents with one or more pre-stored headers corresponding to the identified business type;
extract insurance data corresponding to the one or more matched headers, wherein the matching and extracting is based on the identified file format; and
convert the extracted insurance data into XML format using one or more pre-stored XML elements corresponding to the one or more matched headers.

### Claim 2 (depends on 1)

The system of claim 1, wherein the extraction and conversion module comprises:
an image convertor configured to convert one or more Portable Document Format (PDF) files to one or more images if the identified file format of the one or more submitted documents is PDF;
an image processing module configured to process the one or more images to isolate one or more insurance data blocks, wherein the one or more images are received from at least one of: the image convertor and directly from the one or more users after being identified by the format classification module;
a data extractor configured to extract the insurance data from at least one of:
the isolated one or more insurance data blocks; and
one or more spreadsheets in the form of a two dimensional array if the identified file format of the one or more submitted documents is in a spreadsheet format; and
an XML convertor configured to convert the extracted insurance data using the one or more pre-stored XML elements.

### Claim 3 (depends on 1, 2)

The system of claim 1 and 2, wherein the extraction and conversion module further comprises:
a data cleansing module configured to cleanse the extracted insurance data from the one or more spreadsheets; and
a data structuring module configured to match the cleansed insurance data with the one or more pre-stored headers to identify header row, the one or more headers and the corresponding insurance data.

### Claim 4 (depends on 3)

The system of claim 3, wherein cleansing the extracted insurance data comprises removing at least one of: rows and columns that contain data elements less than a predetermined threshold value, duplicate data and unwanted data.

### Claim 5 (depends on 2)

The system of claim 2, wherein processing the one or more images to isolate the one or more insurance data blocks by the image processing module comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images match with the one or more pre-stored empty templates then the one or more images are considered to be structured;
correcting at least one of: rotation, scale and translation of the one or more structured images to fit the one or more matched empty templates;
matching the one or more headers from the one or more corrected images with the one or more pre-stored headers corresponding to the one or more matched empty templates;
isolating the one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected structured images.

### Claim 6 (depends on 2)

The system of claim 2, wherein processing the one or more images to isolate the one or more insurance data blocks by the image processing module comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images do not match with the one or more pre-stored empty templates then the one or more images are considered to be unstructured;
correcting rotation of the one or more unstructured images;
determining pattern of the one or more corrected unstructured images, wherein the determined pattern is an arrangement of the one or more headers and the corresponding insurance data in the one or more corrected unstructured images;
matching the one or more headers from the one or more corrected unstructured images with the one or more pre-stored headers using the determined pattern; and
isolating one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected unstructured images using the determined pattern.

### Claim 7 (depends on 1)

The system of claim 1 further comprising a submission channels interface configured to facilitate submitting the one or more documents related to insurance directly via one or more submission channels.

### Claim 8 (depends on 7)

The system of claim 7, wherein the one or more submission channels comprise at least one of: a facsimile, an e-mail, an instant message and a digital scanner.

### Claim 9 (depends on 1)

The system of claim 1, wherein the business type refers to type of the one or more submitted documents corresponding to a line of insurance business comprising at least one of: Association for Cooperative Operations Research and Development (ACORD) forms, Cisgem forms, schedules, lossrun reports, Market Value Reduction (MVR) reports and addendums to insurance policy.

### Claim 10 (depends on 2)

The system of claim 2, wherein the one or more images have one or more image file formats comprising at least one of: a Joint Photographic Experts Group (JPEG) format, a JPEG File Interchange Format (JFIF), an Exchangeable Image File Format (EXIF), a Tagged Image File Format (TIFF), a Graphics Interchange Format (GIF), a bitmap image file format, a Portable Network Graphics (PNG) format and any other image file format.

### Claim 11 (depends on 2)

The system of claim 2, wherein the one or more spreadsheets have one or more spreadsheet formats comprising at least one of: a Microsoft® Excel® file format, an Open Document Format (ODF), a Lotus 1-2-3® worksheet format, a Microsoft® Works spreadsheet format, a VisiCalc® spreadsheet format, an Abykus worksheet format, an Ability spreadsheet format and any other spreadsheet format.

### Claim 12 (depends on 1)

The system of claim 1, wherein each of the one or more pre-stored XML elements are XML transformations corresponding to the one or more pre-stored headers.

### Claim 13

A computer-implemented method for extracting insurance data from one or more documents having one or more file formats and converting into Extensible Markup Language (XML) format, via program instructions stored in a memory and executed by a processor, the computer-implemented method comprising:
submitting one or more documents related to insurance;
identifying the one or more submitted documents based on a business type;
identifying file format of the one or more submitted documents;
matching one or more headers in the one or more submitted documents with one or more pre-stored headers corresponding to the identified business type, wherein the matching of one or more headers is based on the identified file format;
extracting insurance data corresponding to the one or more matched headers, wherein the extraction of the insurance data is based on the identified file format; and
converting the extracted insurance data into XML format using one or more pre-stored XML elements corresponding to the one or more matched headers.

### Claim 14 (depends on 13)

The computer-implemented method of claim 13, wherein the method for extracting the insurance data from the one or more documents having one or more file formats and converting into XML format further comprises:
converting one or more PDF files into one or more images if the identified file format of the one or more submitted documents is PDF;
processing the one or more images to isolate one or more insurance data blocks, wherein the one or more images comprise at least one of:
the one or more converted images; and
the one or more submitted documents identified as the one or more images;
extracting the insurance data from the isolated insurance data blocks; and
converting the extracted insurance data using the one or more pre-stored XML elements.

### Claim 15 (depends on 14)

The computer-implemented method of claim 14, wherein the step of processing the one or more images to isolate the one or more insurance data blocks further comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images match with the one or more pre-stored empty templates then the one or more images are considered to be structured;
correcting at least one of: rotation, scale and translation of the one or more structured images to fit the one or more matched empty templates;
matching the one or more headers from the one or more corrected structured images with the one or more pre-stored headers corresponding to the one or more matched empty templates; and
isolating the one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected structured images.

### Claim 16 (depends on 14)

The computer-implemented method of claim 14, wherein the step of processing the one or more images to isolate the one or more insurance data blocks further comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images do not match with the one or more pre-stored empty templates then the one or more images are considered to be unstructured;
correcting rotation of the one or more unstructured images;
determining pattern of the one or more corrected unstructured images, wherein the determined pattern is an arrangement of the one or more headers and the corresponding insurance data in the one or more corrected unstructured images;
matching the one or more headers from the one or more corrected unstructured images with the one or more pre-stored headers using the determined pattern; and
isolating the one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected unstructured images using the determined pattern.

### Claim 17 (depends on 16)

The computer-implemented method of claim 16, wherein the step of correcting the rotation of the one or more unstructured images comprises:
processing the one or more unstructured images using image thresholding to generate one or more binary images, wherein the image thresholding facilitates image segmentation by converting the one or more unstructured images in the form of one or more gray scale images into the one or more binary images;
dilating the one or more binary images using image dilation to generate one or more dilated images, wherein dilating the one or more binary images facilitates in expanding objects and shapes in the one or more binary images;
processing the one or more dilated images using image thinning to generate the one or more thinned images, wherein thinning the one or more dilated images facilitates image skeletonization;
estimating a skew angle of the one or more thinned images using Hough transform; and
correcting the rotation of the one or more thinned images based on the estimated skew angle to generate the one or more corrected unstructured images.

### Claim 18 (depends on 13)

The computer-implemented method of claim 13, wherein the method for extracting the insurance data from the one or more documents having one or more file formats and converting into XML format further comprises:
extracting the insurance data from one or more spreadsheets in the form of a two dimensional array if the identified file format of the one or more submitted documents is in a spreadsheet format;
cleansing the extracted insurance data, wherein the extracted insurance data is cleansed by removing at least one of: rows and columns that contain data elements less than a predetermined threshold value, duplicate data and unwanted data;
matching the cleansed insurance data with the one or more pre-stored headers to identify header row, the one or more headers and the corresponding insurance data;
matching the one or more headers in the header row with the one or more pre-stored headers, wherein if the one or more headers in the header row do not match with the one or more pre-stored headers then the one or more unmatched headers are renamed; and
converting the insurance data corresponding to at least one of: the one or more matched headers and the one or more renamed headers into XML format using the one or more pre-stored XML elements corresponding to the one or more matched headers and the one or more renamed headers respectively.

### Claim 19 (depends on 13)

The computer-implemented method of claim 13, wherein the one or more documents are submitted via at least one of: a user interface and a submission channels interface.

### Claim 20

A computer program product for extracting insurance data from one or more documents having one or more file formats and converting into Extensible Markup Language (XML) format, the computer program product comprising:
a non-transitory computer-readable medium having computer-readable program code stored thereon, the computer-readable program code comprising instructions that when executed by a processor, cause the processor to:
submit one or more documents related to insurance;
identify the one or more submitted documents based on a business type;
identify file format of the one or more submitted documents;
match one or more headers in the one or more submitted documents with one or more pre-stored headers corresponding to the identified business type, wherein the matching of one or more headers is based on the identified file format;
extract insurance data corresponding to the one or more matched headers, wherein the extraction of the insurance data is based on the identified file format; and
convert the extracted insurance data into XML format using one or more pre-stored XML elements corresponding to the one or more matched headers.

### Claim 21 (depends on 20)

The computer program product of claim 20, wherein extracting the insurance data from the one or more documents having one or more file formats and converting into XML format further comprises:
converting one or more PDF files into one or more images if the identified file format of the one or more submitted documents is PDF;
processing the one or more images to isolate one or more insurance data blocks, wherein the one or more images comprise at least one of:
the one or more converted images; and
the one or more submitted documents identified as the one or more images;
extracting the insurance data from the isolated insurance data blocks; and
converting the extracted insurance data using the one or more pre-stored XML elements.

### Claim 22 (depends on 21)

The computer program product of claim 21, wherein processing the one or more images to isolate the one or more insurance data blocks further comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images match with the one or more pre-stored empty templates then the one or more images are considered to be structured;
correcting at least one of: rotation, scale and translation of the one or more structured images to fit the one or more matched empty templates;
matching the one or more headers from the one or more corrected structured images with the one or more pre-stored headers corresponding to the one or more matched empty templates; and
isolating the one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected structured images.

### Claim 23 (depends on 21)

The computer program product of claim 21, wherein processing the one or more images to isolate the one or more insurance data blocks further comprises:
determining structure of the one or more images by matching the one or more images with one or more pre-stored empty templates corresponding to the identified business type, wherein if the one or more images do not match with the one or more pre-stored empty templates then the one or more images are considered to be unstructured;
correcting rotation of the one or more unstructured images;
determining pattern of the one or more corrected unstructured images, wherein the determined pattern is an arrangement of the one or more headers and the corresponding insurance data in the one or more corrected unstructured images;
matching the one or more headers from the one or more corrected unstructured images with the one or more pre-stored headers using the determined pattern; and
isolating the one or more insurance data blocks corresponding to the one or more matched headers from the one or more corrected unstructured images using the determined pattern.

### Claim 24 (depends on 23)

The computer program product of claim 23, wherein correcting the rotation of the one or more unstructured images comprises:
processing the one or more unstructured images using image thresholding to generate one or more binary images, wherein the image thresholding facilitates image segmentation by converting the one or more unstructured images in the form of one or more gray scale images into the one or more binary images;
dilating the one or more binary images using image dilation to generate one or more dilated images, wherein dilating the one or more binary images facilitates in expanding objects and shapes in the one or more binary images;
processing the one or more dilated images using image thinning to generate the one or more thinned images, wherein thinning the one or more dilated images facilitates image skeletonization;
estimating a skew angle of the one or more thinned images using Hough transform; and
correcting the rotation of the one or more thinned images based on the estimated skew angle to generate the one or more corrected unstructured images.

### Claim 25 (depends on 20)

The computer program product of claim 20, wherein extracting the insurance data from the one or more documents having one or more file formats and converting into XML format further comprises:
extracting the insurance data from one or more spreadsheets in the form of a two dimensional array if the identified file format of the one or more submitted documents is in a spreadsheet format;
cleansing the extracted insurance data, wherein the extracted insurance data is cleansed by removing at least one of: rows and columns that contain data elements less than a predetermined threshold value, duplicate data and unwanted data;
matching the cleansed insurance data with the one or more pre-stored headers to identify header row, the one or more headers and the corresponding insurance data;
matching the one or more headers in the header row with the one or more pre-stored headers, wherein if the one or more headers in the header row do not match with the one or more pre-stored headers then the one or more unmatched headers are renamed; and
converting the insurance data corresponding to at least one of: the one or more matched headers and the one or more renamed headers into XML format using the one or more pre-stored XML elements corresponding to the one or more matched headers and the one or more renamed headers respectively.