Method and system for bootstrapping an OCR engine for license plate recognition

• US 9,501,707 B2
• Filed: 04/16/2015
• Issued: 11/22/2016
• Est. Priority Date: 04/16/2015
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Bulan, Orhan, Fillion, Claude, Burry, Aaron M., Kozitsky, Vladimir
• Primary Examiner(s): FUJITA, KATRINA R
• Application Number:
  US14/688,255
• Publication Number:
  US 20160307061A1
• Time in Patent Office: 586 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 30/10 Character recognition

Abstract

Methods and systems for bootstrapping an OCR engine for license plate recognition. One or more OCR engines can be trained utilizing purely synthetically generated characters. A subset of classifiers, which require augmentation with real examples, along how many real examples are required for each, can be identified. The OCR engine can then be deployed to the field with constraints on automation based on this analysis to operate in a “bootstrapping” period wherein some characters are automatically recognized while others are sent for human review. The previously determined number of real examples required for augmenting the subset of classifiers can be collected. Each subset of identified classifiers can then be retrained as the number of real examples required becomes available.

Claims

1. A method for optimizing an ALPR system, said method comprising:

generating synthetic images with respect to each character on a license plate image;

training at least one classifier utilizing said synthetic images;

determining a number of samples of real images required for said each character based on attributes of said at least one classifier training utilizing said synthetic images; and

retraining said at least one classifier utilizing said synthetic images and said real images as said real images become available.

2. The method of claim 1 wherein said at least one classifier comprises a one-vs-all classifier.

3. The method of claim 1 wherein said at least one classifier comprises an OCR engine.

4. The method of claim 3 further comprising:

deploying said OCR engine with constraints based on said training and retraining of said at least one classifier; and

operating said OCR engine in a bootstrapping period wherein some characters are automatically recognized while other characters are transmitted for human review.

5. The method of claim 1 wherein determining a number of samples of real images, further comprises:

identifying a subset of said at least one classifier requiring augmentation with said real images and a number of real images required for each subset of said at least one classifier.

6. The method of claim 5 further comprising:

collecting a previously determined number of said real images required for augmenting said each subset of said at least one classifier; and

retraining said each subset of identified classifiers among said at least one classifier as a number of real examples of said real images required becomes available.

7. The method of claim 6 wherein said at least one classifier comprises a one-vs-all classifier.

8. The method of claim 6 wherein said at least one classifier comprises an OCR engine.

9. A system for optimizing an ALPR system, said system comprising:

at least one processor; and

a computer-usable medium embodying computer program code, said computer-usable medium capable of communicating with said at least one processor, said computer program code comprising instructions executable by said at least one processor and configured for;

generating synthetic images with respect to each character on a license plate image;

training at least one classifier utilizing said synthetic images;

determining a number of samples of real images required for said each character based on attributes of said at least one classifier training utilizing said synthetic images; and

retraining said at least one classifier utilizing said synthetic images and said real images as said real images become available.

10. The system of claim 9 wherein said at least one classifier comprises a one-vs-all classifier.

11. The system of claim 9 wherein said at least one classifier comprises an OCR engine.

12. The system of claim 9 wherein said instructions determining a number of samples of real mages, further comprise instructions for:

identifying a subset of said at least one classifier requiring augmentation with said real images and a number of real images required for each subset of said at least one classifier.

13. The system of claim 12 wherein said instructions are further configured for:

deploying said OCR engine with constraints based on said training and retraining of said at least one classifier; and

operating said OCR engine in a bootstrapping period wherein some characters are automatically recognized while other characters are transmitted for human review.

14. The system of claim 13 wherein said instructions are further configured for:

collecting a previously determined number of said real images required for augmenting said each subset of said at least one classifier; and

retraining said each subset of identified classifiers among said at least one classifier as a number of real examples of said real images required becomes available.

15. The system of claim 14 wherein said at least one classifier comprises a one-vs-all classifier.

16. The system of claim 14 wherein said at least one classifier comprises an OCR engine.

17. The system of claim 14 wherein said at least one classifier comprises at least one of:

a one-vs-all classifier and an OCR engine.

18. A system for optimizing an ALPR system, said system comprising:

at least one imaging capturing unit that captures a license plate image;

at least one processor that communicates electronically with said at least one image capturing unit; and

a computer-usable medium embodying computer program code, said computer-usable medium capable of communicating with said at least one processor, said computer program code comprising instructions executable by said at least one processor and configured for;

generating synthetic images with respect to each character on said license plate image;

training at least one classifier utilizing said synthetic images;

determining a number of samples of real images required for said each character based on attributes of said at least one classifier training utilizing said synthetic images; and

retraining said at least one classifier utilizing said synthetic images and said real images as said real images become available.

19. The system of claim 18 wherein said instructions determining a number of samples of real images, further comprise instructions for:

identifying a subset of said at least one classifier requiring augmentation with said real images and a number of real images required for each subset of said at least one classifier.

20. The system of claim 18 wherein said instructions are further configured for:

deploying said OCR engine with constraints based on said training and retraining of said at least one classifier;

operating said OCR engine in a bootstrapping period wherein some characters are automatically recognized while other characters are transmitted for human review;

collecting a previously determined number of said real images required for augmenting said each subset of said at least one classifier; and

retraining said each subset of identified classifiers among said at least one classifier as a number of real examples of said real images required becomes available.