Text extraction using optical character recognition

Patent metadata

Field	Value
Publication number	US-11961316-B2
Application number	US-202217741113-A
Country	US
Kind code	B2
Filing date	May 10, 2022
Priority date	May 10, 2022
Publication date	Apr 16, 2024
Grant date	Apr 16, 2024

Abstract

Provided herein are systems and methods for extracting text from a document. Different optical character recognition (OCR) tools are used to extract different versions of the text in the document. Metrics evaluating the quality of the extracted text are compared to identify and select higher quality extracted text. A selected portion of text is compared to a threshold to ensure minimal quality. The selected portion of text is then saved. Error correction can be applied to the selected portion of text based on errors specific to the OCR tools or the document contents.

First claim

What is claimed is:

1. A method, comprising:
extracting, by at least one processor, a first set of text from a document using a first optical character recognition (OCR) tool;
extracting, by the at least one processor, a second set of text from the document using a second OCR tool;
comparing, by the at least one processor, a first metric of the first set of text to a second metric of the second set of text, the first metric and the second metric measuring a respective document-level OCR quality of the first set of text and the second set of text;
selecting, by the at least one processor, a portion of text from between the first set of text and the second set of text based on the first metric and the second metric; and
storing, by the at least one processor, the selected portion of text in a searchable format in response to the portion of text having a third metric greater than a threshold, the third metric measuring a page-level OCR quality of the portion of text.

2. The method of claim 1 , wherein the first OCR tool extracts the first set of text from the document by:
converting, by the at least one processor, the document into a set of images, each page of the document being converted into a respective image from the set of images; and
extracting, by the at least one processor, the first set of text from at least one of the respective images from the set of images.

3. The method of claim 1 , further comprising:
identifying, by the at least one processor, a group of characters in the selected portion of text comprising an erroneous character, wherein the erroneous character is expected to be a different character based on the group of characters;
correcting, by the at least one processor, the selected portion of text by changing the erroneous character to the different character.

4. The method of claim 1 , further comprising:
identifying, by the at least one processor, a two or more characters in the selected portion of text that is missing a space between two characters in the two or more characters based on the two or more characters;
correcting, by the at least one processor, the selected portion of text by adding the space between the two characters in the two or more characters.

5. The method of claim 1 , further comprising:
identifying, by the at least one processor, a superscript character or a subscript character in a group of characters in the selected portion of text that is unexpected based on the group of characters;
correcting, by the at least one processor, the selected portion of text by deleting the superscript character or the subscript character from the group of characters.

6. The method of claim 1 , further comprising determining the first metric and the second metric, wherein the first metric and the second metric comprise a respective number of words in the first set of text and the second set of text extracted from the document.

7. The method of claim 1 , further comprising, determining, for the selected set of text, the third metric, wherein the third metric is a respective number of words in the selected set of text extracted from the document divided by a number of pages in the document.

8. A system, comprising:
one or more processors;
memory communicatively coupled to the one or more processors, the memory storing instructions which, when executed by the one or more processors, cause the one or more processors to:
extracting a first set of text from a document using a first optical character recognition (OCR) tool;
extracting, by the at least one processor, a second set of text from the document using a second OCR tool;
comparing a first metric of the first set of text to a second metric of the second set of text, the first metric and the second metric measuring a respective document-level OCR quality of the first set of text and the second set of text;
selecting a portion of text from between the first set of text and the second set of text based on the first metric and the second metric; and
storing the selected portion of text in a searchable format in response to the selected portion of text having a third metric greater than a threshold, the third metric measuring a page-level OCR quality of the selected portion of text.

9. The system of claim 8 , the instructions further configured to implement the first OCR tool, wherein the first OCR tool is configured to extract the first set of text from the document by:
converting the document into a set of images, each page of the document being converted into a respective image from the set of images; and
extracting the first set of text from at least one of the respective images from the set of images.

10. The system of claim 8 , wherein the instructions further cause the one or more processors to:
identify a group of characters in the selected portion of text comprising an erroneous character, wherein the erroneous character is expected to be a different character based on the group of characters;
correct the selected portion of text by changing the erroneous character to the different character.

11. The system of claim 8 , wherein the instructions further cause the one or more processors to:
identify a two or more characters in the selected portion of text that is missing a space between two characters in the two or more characters based on the two or more characters;
correct the selected portion of text by adding the space between the two characters in the two or more characters.

12. The system of claim 8 , wherein the instructions further cause the one or more processors to:
identify a superscript character or a subscript character in a group of characters in the selected portion of text that is unexpected based on the group of characters;
correct the selected portion of text by deleting the superscript character or the subscript character from the group of characters.

13. The system of claim 8 , wherein the instructions further cause the one or more processors to determine, the first metric and the second metric, wherein the first metric and the second metric comprise a respective number of words in the first set of text and the second set of text extracted from the document.

14. The system of claim 8 , wherein the instructions further cause the one or more processors to determine, for the selected portion of text, the third metric, wherein the third metric is a respective number of words in the selected portion of text extracted from the document divided by a number of pages in the document.

15. A non-transitory computer-readable storage medium having computer-readable code thereon, the non-transitory computer-readable storage medium including instructions configured to cause a computer system to perform operations, comprising:
extracting a first set of text from a document using a first optical character recognition (OCR) tool;
extracting, by the at least one processor, a second set of text from the document using a second OCR tool;
comparing a first metric of the first set of text to a second metric of the second set of text, the first metric and the second metric measuring a respective document-level OCR quality of the first set of text and the second set of text;
selecting a portion of text from between the first set of text and the second set of text based on the first metric and the second metric; and
storing the selected portion of text in a searchable format in response to the selected portion of text having a third metric greater than a threshold, the third metric measuring a page-level OCR quality of the selected portion of text.

16. The non-transitory computer-

Assignees

• Capital One Services Llc

Inventors

• Demchalk Chris
• Parker Ryan M
• Kumar Lokesh Vijay
• Fromknecht Brian

Classifications

• G06V30/133 Primary

  Evaluation of quality of the acquired characters · CPC title

• G06V30/155

  Removing patterns interfering with the pattern to be recognised, such as ruled lines or underlines · CPC title

• G06V30/26

  Techniques for post-processing, e.g. correcting the recognition result · CPC title

• G06V30/41

  Analysis of document content (recognition of printed characters based on code marks G06V30/224) · CPC title