Determining a class associated with an image

• US 8,873,856 B1
• Filed: 09/09/2013
• Issued: 10/28/2014
• Est. Priority Date: 07/14/2009
• Status: Active Grant

• Current Assignee: Matrox Electronic Systems Ltd.
• Original Assignee: Matrox Electronic Systems Ltd.
• Inventors: Chapleau, Sylvain, Paquin, Vincent
• Primary Examiner(s): SHAH, UTPAL D
• Application Number:
  US14/020,934
• Time in Patent Office: 414 Days
• Field of Search: 382/177, 382/321, 382/224
• US Class Current: 382/173
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/153 using recognition of characters or words

Abstract

The technology is directed to determining a class associated with an image. In some examples, a method determines the class associated with an image. The method can include determining a segmentation score for an image segment based on a comparison of the image segment and a region of an image. The region of the image can be associated with the image segment. The method further includes determining a confidence score for the image segment based on the segmentation score and a classification score. The classification score can be indicative of a similarity between the image segment and at least one class. The method further includes determining a class associated with the image based on the confidence score. The method further includes outputting the class associated with the image.

Claims

1. A method for determining a class associated with an image, the method comprising:

generating at least one image segment based on a segmentation of the image;

determining a segmentation score for an image segment included in the at least one image segment based on a comparison of the image segment and a region of the image, the region of the image associated with the image segment;

determining a confidence score for the image segment based on the segmentation score and a classification score, the classification score indicative of a similarity between the image segment and at least one class;

determining a class associated with the image based on the confidence score; and

outputting the class associated with the image.

2. The method of claim 1, wherein the at least one image segment includes a plurality of image segments, wherein the image segment is a first image segment, wherein the confidence score is a first confidence score, wherein the region of the image is a first region of the image and wherein the method further comprises:

determining a second segmentation score for a second image segment included in the plurality of image segments based on a comparison of the second image segment and a second region of the image, the second region of the image associated with the second image segment; and

determining a second confidence score for the second image segment based on the second segmentation score and a second classification score, the second classification score indicative of a similarity between the second image segment and the at least one class.

3. The method of claim 2, wherein the determining the class associated with the image further comprises determining the class associated with the image based on the first confidence score and the second confidence score.

4. The method of claim 2, further comprising generating the first image segment and the second image segment based on a first segmentation and a second segmentation, respectively, of substantially the same region of the image.

5. The method of claim 2, further comprising generating the first image segment and the second image segment based on a segmentation of regions of the image substantially distinct from one another.

6. The method of claim 1, further comprising comparing pixels of the image segment with pixels of the region of the image to form the comparison of the image segment and the region of the image.

7. The method of claim 1, further comprising determining the classification score indicative of the similarity between the image segment and the at least one class.

8. The method of claim 1, further comprising determining the at least one class based on a comparison of the image segment with a plurality of classes, each of the plurality of classes associated with a character.

9. The method of claim 1, further comprising:

receiving the image from an image acquisition device;

binarizing the image to form a binarized image; and

segmenting the binarized image based on objects within the binarized image to form the image segment.

10. The method of claim 1, wherein the determining the class associated with the image further comprising determining the class associated with the image based on the confidence score by comparing the confidence score to a threshold value.

11. An apparatus comprising:

a camera configured to acquire an image;

at least one processor coupled to the camera; and

a computer storage medium storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations including;

generating at least one image segment based on a segmentation of the image;

determining a segmentation score for an image segment included in the at least one image segment based on a comparison of the image segment and a region of the image, the region of the image associated with the image segment;

determining a confidence score for the image segment based on the segmentation score and a classification score, the classification score indicative of a similarity between the image segment and at least one class;

determining a class associated with the image based on the confidence score; and

outputting the class associated with the image.

12. A system for determining a class associated with an image, the system comprising:

at least one processor;

a segmentation module configured to employ the at least one processor to generate at least one image segment based on a segmentation of the image;

a segmentation score module configured to employ the at least one processor to determine a segmentation score for an image segment included in the at least one image segment based on a comparison of the image segment and a region of the image, the region of the image associated with the image segment;

a confidence score module configured to employ the at least one processor to determine a confidence score for the image segment based on the segmentation score and a classification score, the classification score indicative of a similarity between the image segment and at least one class;

a class determination module configured to employ the at least one processor to determine a class associated with the image based on the confidence score; and

a class output module configured to employ the at least one processor to output the class associated with the image.

13. The system of claim 12, wherein the at least one image segment includes a plurality of image segments, wherein the image segment is a first image segment, wherein the confidence score is a first confidence score, wherein the region of the image is a first region of the image, and wherein the system further comprises:

the segmentation score module further configured to employ the at least one processor to determine a second segmentation score for a second image segment included in the plurality of image segments based on a comparison of the second image segment and a second region of the image, the second region of the image associated with the second image segment; and

the confidence score module further configured to employ the at least one processor to determine a second confidence score for the second image segment based on the second segmentation score and a second classification score, the second classification score indicative of a similarity between the second image segment and the at least one class.

14. The system of claim 13, further comprising:

the class determination module further configured to employ the at least one processor to determine the class associated with the image based on the first confidence score and the second confidence score.

15. The system of claim 13, further comprising the segmentation module further configured to employ the at least one processor to generate the first image segment and the second image segment based on a first segmentation and a second segmentation, respectively, of substantially the same region of the image.

16. The system of claim 13, further comprising the segmentation module further configured to employ the at least one processor to generate the first image segment and the second image segment based on a segmentation of regions of the image substantially distinct from one another.

17. The system of claim 12, further comprising the segmentation score module further configured to employ the at least one processor to compare pixels of the image segment with pixels of the region of the image to form the comparison of the image segment and the region of the image.

18. The system of claim 12, further comprising a classification score module configured to employ the at least one processor to determine the classification score indicative of the similarity between the image segment and the at least one class.

19. The system of claim 12, further comprising a classification score module configured to employ the at least one processor to determine the at least one class based on a comparison of the image segment with a plurality of classes, each of the plurality of classes associated with a character.

20. The system of claim 12, further comprising:

an image acquisition module configured to employ the at least one processor to receive the image,

wherein the segmentation module is further configured to employ the at least one processor to;

binarize the image to form a binarized image; and

segment the binarized image based on objects within the binarized image to form the image segment.

21. The system of claim 12, further comprising the class determination module further configured to employ the at least one processor to determine the class associated with the image based on the confidence score by comparing the confidence score to a threshold value.