Granted Patent B2

US 12,260,659 · App. 17/660,639 · Granted Mar 25, 2025

Inventors: Ophir Azulai (Tivon, IL); Daniel Nechemia Rotman (Haifa, IL); Udi Barzelay (Haifa, IL)

Assignee: International Business Machines Corporation

G06V30/245G06F40/30G06V30/1444G06V30/153

Abstract

Claims (57)

1. A computer-implemented method comprising:

receiving a document having different font attributes amongst a plurality of words respectively comprised of at least one character;

generating a dense image document from the document by:

setting the plurality of words to a predefined size;

removing blank spaces from the document; and

altering an order of characters relative to the document by placing the characters in a random order relative to the document;

determining characteristics of the characters in the dense image document;

aggregating the characteristics for at least one word; and

annotating the at least one word with a font attribute based on the aggregated characteristics.

2. The method of claim 1 , wherein the different font attributes are selected from a group consisting of: bold, italic, and underline.

3. The method of claim 1 , wherein receiving the document further comprises:

applying bounding boxes to respective words of the plurality of words in the document; and

applying bounding boxes to respective characters in the respective words of the plurality of words.

4. The method of claim 3 , wherein setting the plurality of words to the predefined size is performed by modifying a height the plurality of words while maintaining an aspect ratio of bounding boxes corresponding to the respective words of the plurality of words.

5. The method of claim 1 , wherein placing the characters in the random order causes at least one character with the font attribute and in a middle of the at least one word to be adjacent to a character without the font attribute in the dense image document.

6. The method of claim 1 , wherein determining the characteristics of the characters in the dense image document further comprises:

inputting the dense image document to a semantic segmentation model; and

receiving the characteristics of the characters as an output from the semantic segmentation model.

7. The method of claim 1 , wherein the characteristics of the characters comprise a ratio of pixels divided by a text area.

8. The method of claim 1 , wherein aggregating the characteristics for the at least one word comprises averaging the characteristics of a set of characters corresponding to the at least one word.

9. The method of claim 1 , wherein annotating the at least one word with the font attribute includes generating an annotated document, and wherein the method further comprises:

performing natural language processing (NLP) on the annotated document.

10. The method of claim 1 , wherein the method is performed by one or more computers according to software that is downloaded to the one or more computers from a remote data processing system, and wherein the method further comprises:

metering a usage of the software; and

generating an invoice based on metering the usage.

11. A system comprising:

one or more computer readable storage media storing program instructions; and

one or more processors which, in response to executing the program instructions, are configured to perform a method comprising:

receiving a document having different font attributes amongst a plurality of words respectively comprised of at least one character;

generating a dense image document from the document by:

setting the plurality of words to a predefined size;

removing blank spaces from the document; and

altering an order of characters relative to the document by placing the characters in a random order relative to the document;

determining characteristics of the characters in the dense image document;

aggregating the characteristics for at least one word; and

annotating the at least one word with a font attribute based on the aggregated characteristics.

12. The system of claim 11 , wherein the different font attributes are selected from a group consisting of: bold, italic, and underline.

13. The system of claim 11 , wherein placing the characters in the random order causes at least one character with the font attribute and in a middle of the at least one word to be adjacent to a character without the font attribute in the dense image document.

14. The system of claim 11 , wherein determining the characteristics of the characters in the dense image document further comprises:

inputting the dense image document to a semantic segmentation model; and

receiving the characteristics of the characters as an output from the semantic segmentation model.

15. The system of claim 11 , wherein the characteristics of the characters comprise a ratio of pixels divided by a text area.

16. A computer program product embodied on one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising instructions configured to cause one or more processors to perform a method comprising:

receiving a document having different font attributes amongst a plurality of words respectively comprised of at least one character;

generating a dense image document from the document by:

setting the plurality of words to a predefined size;

removing blank spaces from the document; and

altering an order of characters relative to the document by placing the characters in a random order relative to the document;

determining characteristics of the characters in the dense image document;

aggregating the characteristics for at least one word; and

annotating the at least one word with a font attribute based on the aggregated characteristics.

17. The computer program product of claim 16 , wherein the different font attributes are selected from a group consisting of: bold, italic, and underline.

18. The computer program product of claim 16 , wherein placing the characters in the random order causes at least one character with the font attribute and in a middle of the at least one word to be adjacent to a character without the font attribute in the dense image document.

19. The computer program product of claim 16 , wherein determining the characteristics of the characters in the dense image document further comprises:

inputting the dense image document to a semantic segmentation model; and

receiving the characteristics of the characters as an output from the semantic segmentation model.

20. The computer program product of claim 16 , wherein the characteristics of the characters comprise a ratio of pixels divided by a text area.

Assignments (1)

Continuity (1)

References Cited (20)