Leveraging character-by-character image classifiers to improve license plate state identification

• US 9,405,985 B1
• Filed: 03/17/2015
• Issued: 08/02/2016
• Est. Priority Date: 03/17/2015
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Burry, Aaron M., Kozitsky, Vladimir
• Primary Examiner(s): Abdi, Amara
• Application Number:
  US14/660,255
• Time in Patent Office: 504 Days
• US Class Current: 1/1
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/245 Font recognition
• CPC Class Codes: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

Methods and systems for enhancing the accuracy of license plate state identification in an ALPR (Automated License Plate Recognition) system. This is accomplished through use of individual character-by-character image-based classifiers that are trained to distinguish between the fonts for different states. At runtime, the OCR result for the license plate code can be used to determine which character in the plate would provide the highest discriminatory power for arbitrating between candidate state results. This classifier is then applied to the individual character image to provide a final selection of the estimated state/jurisdiction for the plate.

Claims

1. A method for enhancing the accuracy of license plate state identification, said method comprising:

training at least one character-by-character image-based classifier that differentiates between at least two possible different states including, distinguishing between fonts on license plate images with respect to said at least two possible different states;

determining at least one character among a plurality of characters on a license plate image that provides a high discriminatory power for arbitrating between candidate state results with respect to said at least two possible different states; and

applying said at least one character-by-character image-based classifier to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image.

2. The method of claim 1 further comprising employing an OCR result to determine which character on said license plate image provides said high discriminatory power.

3. The method of claim 1 further comprising capturing said license plate image via an image-capturing device.

4. The method of claim 3 wherein said image-capturing device comprises an ALPR camera.

5. The method of claim 1 wherein applying said at least one character-by-character image-based classifier to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image, further comprises:

segmenting said at least one character; and

running said at least one character-by-character image-based classifier with respect to said at least one character after said at least one character is segmented.

6. The method of claim 1 wherein said at least one character-by-character image-based classifier comprises a pairwise classifier based on binary classification.

7. The method of claim 1 wherein said at least one character-by-character image-based classifier comprises a multi-class classifier.

8. The method of claim 2 wherein applying said at least one character-by-character image-based classifier to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image, further comprises:

segmenting said at least one character; and

running said at least one character-by-character image-based classifier with respect to said at least one character after said at least one character is segmented, said running of said at least one character-by-character image-based classifier further comprises scanning a table of classifier accuracies.

9. The method of claim 8 wherein said at least one character-by-character image-based classifier comprises a pairwise classifier based on binary classification, wherein said table of classifier accuracies is learned offline.

10. The method of claim 8 wherein said at least one character-by-character image-based classifier comprises a multi-class classifier, wherein said table of classifier accuracies is learned offline.

11. A system for enhancing the accuracy of license plate state identification, 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;

training at least one character-by-character image-based classifier that differentiates between at least two possible different states including distinguishing between fonts on license plate images with respect to said at least two possible different states;

determining at least one character among a plurality of characters on a license plate image that provides a high discriminatory power for arbitrating between candidate state results with respect to said at least two possible different states; and

applying said at least one character-by-character image-based classifier to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image.

12. The system of claim 11 wherein said instructions are further configured for employing an OCR result to determine which character on said license plate image provides said high discriminatory power.

13. The system of claim 11 further comprising capturing said license plate image via an image-capturing device.

14. The system of claim 13 wherein said image-capturing device comprises an ALPR camera.

15. The system of claim 11 wherein said instructions for applying said at least one character-by-character image-based classifier to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image, further comprise instructions configured for:

segmenting said at least one character; and

running said at least one character-by-character image-based classifier with respect to said at least one character after said at least one character is segmented, wherein said running of said at least one character-by-character image-based classifier further comprises scanning a table of classifier accuracies.

16. The system of claim 11 wherein said at least one character-by-character image-based classifier comprises a pairwise classifier based on binary classification.

17. The system of claim 11 wherein said at least one character-by-character image-based classifier comprises a multi-class classifier.

19. The system of claim 17 wherein an OCR result is employed to determine which character on said license plate image provides said high discriminatory power.

18. A system for enhancing the accuracy of license plate state identification, said system comprising:

at least one camera;

at least one character-by-character image-based classifier that communicates with said at least one camera and differentiates between at least two possible different states including distinguishing between fonts on license plate images with respect to said at least two possible different states;

wherein at least one character is determined among a plurality of characters on a license plate image captured by said a least one camera that provides a high discriminatory power for arbitrating between candidate state results with respect to said at least two possible different states; and

wherein said at least one character-by-character image-based classifier is applied to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image.

20. The method of claim 18 wherein said at least one character-by-character image-based classifier is applied to said at least one character to provide a final selection among said candidate state results of an estimated state or jurisdiction associated with said license plate image by:

segmenting said at least one character; and

running said at least one character-by-character image-based classifier with respect to said at least one character after said at least one character is segmented, wherein said running of said at least one character-by-character image-based classifier further comprises scanning a table of classifier accuracies.