Method and System for Optimizing Accuracy-Specificity Trade-offs in Large Scale Visual Recognition

• US 20140086497A1
• Filed: 03/15/2013
• Published: 03/27/2014
• Est. Priority Date: 06/14/2012
• Status: Active Grant

• Current Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Original Assignee: Board of Trustees of the Leland Stanford Junior University (Stanford Management Co.)
• Inventors: Fei-Fei, Li, Deng, Jia, Jonothan, Krause, Alexander, Berg C.
• Granted Patent: US 9,158,965 B2
• Time in Patent Office: Days
• US Class Current: 382/224
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 20/10 Terrestrial scenes scenes under surveillance with static cameras G06V20/52; scenes perceived from the exterior of a vehicle G06V20/56; scenes perceived from the interior of a vehicle G06V20/59
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

As visual recognition scales up to ever larger numbers of categories, maintaining high accuracy is increasingly difficult. Embodiment of the present invention include methods for optimizing accuracy-specificity trade-offs in large scale recognition where object categories form a semantic hierarchy consisting of many levels of abstraction.

Claims

1. A computerized method for classifying images, comprising:

receiving an image hierarchy;

receiving a first image of interest wherein the image includes at least one feature;

classifying at least one feature of the image according to the image hierarchy;

generating a measure of uncertainty associated with the classification of the at least one feature;

optimizing the classification of the at least one feature of the image using the measure of uncertainty; and

generating at least one optimized classification of the at least one feature of the image for the first image of interest.

2. The method of claim 1, wherein optimizing the classification is performed by trading off specificity in the classification for accuracy of the classification.

3. The method of claim 2, wherein the measure of uncertainty exceeds a predetermined level.

4. The method of claim 1, wherein the optimization is performed by implementing a reward algorithm for preferred classifications.

5. The method of claim 4, wherein the reward is implemented to produce classifications with increased specificity.

6. The method of claim 4, wherein the reward includes a measure of information gain.

7. The method of claim 4, wherein the reward includes a measure of decrease in uncertainty.

8. The method of claim 4, wherein the reward algorithm allocates predetermined values in the image hierarchy.

9. The method of claim 1, wherein generating at least one optimized classification of the at least one feature of the image for the first image of interest includes generating at least two optimized classifications for the image of interest.

10. The method of claim 1, wherein optimizing the classification is performed using a Lagrange algorithm.

11. A computer-readable medium including instructions that, when executed by a processing unit, cause the processing unit to implement a method for classifying images, by performing the steps of:

receiving an image hierarchy;

receiving a first image of interest wherein the image includes at least one feature;

classifying at least one feature of the image according to the image hierarchy;

generating a measure of uncertainty associated with the classification of the at least one feature;

optimizing the classification of the at least one feature of the image using the measure of uncertainty; and

generating at least one optimized classification of the at least one feature of the image for the first image of interest.

12. The computer-readable medium of claim 11, wherein optimizing the classification is performed by trading off specificity in the classification for accuracy of the classification.

13. The computer-readable medium of claim 12, wherein the measure of uncertainty exceeds a predetermined level.

14. The computer-readable medium of claim 11, wherein the optimization is performed by implementing a reward algorithm for preferred classifications.

15. The computer-readable medium of claim 14, wherein the reward is implemented to produce classifications with increased specificity.

16. The computer-readable medium of claim 14, wherein the reward includes a measure of information gain.

17. The computer-readable medium of claim 14, wherein the reward includes a measure of decrease in uncertainty.

18. The computer-readable medium of claim 14, wherein the reward algorithm allocates predetermined values in the image hierarchy.

19. The computer-readable medium of claim 11, wherein generating at least one optimized classification of the at least one feature of the image for the first image of interest includes generating at least two optimized classifications for the image of interest.

20. The computer-readable medium of claim 11, wherein optimizing the classification is performed using a Lagrange algorithm.

21. A computing device comprising:

a data bus;

a memory unit coupled to the data bus;

at least one processing unit coupled to the data bus and configured to receive an image hierarchy;

receive a first image of interest wherein the image includes at least one feature;

classify at least one feature of the image according to the image hierarchy;

generate a measure of uncertainty associated with the classification of the at least one feature;

optimize the classification of the at least one feature of the image using the measure of uncertainty; and

generate at least one optimized classification of the at least one feature of the image for the first image of interest.