Video Object Classification

• US 20100054535A1
• Filed: 08/28/2008
• Published: 03/04/2010
• Est. Priority Date: 08/28/2008
• Status: Active Grant

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Chen, Longbin, Hampapur, Arun, Zhai, Yun, Brown, Lisa Marie, Feris, Rogerio Schmidt
• Granted Patent: US 8,249,301 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06V 10/235 based on user input or interaction
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G06V 2201/08 Detecting or categorising vehicles
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

Techniques for classifying one or more objects in at least one video, wherein the at least one video comprises a plurality of frames are provided. One or more objects in the plurality of frames are tracked. A level of deformation is computed for each of the one or more tracked objects in accordance with at least one change in a plurality of histograms of oriented gradients for a corresponding tracked object. Each of the one or more tracked objects is classified in accordance with the computed level of deformation.

Claims

1. A method for classifying one or more objects in at least one video, wherein the at least one video comprises a plurality of frames, the method comprising:

tracking one or more objects in the plurality of frames;

computing a level of deformation for each of the one or more tracked objects in accordance with at least one change in a plurality of histograms of oriented gradients for a corresponding tracked object; and

classifying each of the one or more tracked objects in accordance with the computed level of deformation.

2. The method of claim 1, wherein a given tracked object is classified as at least one of a person and a vehicle.

3. The method of claim 1, further comprising the step of computing at least one of a size vector for each of the one or more tracked objects, a velocity vector for each of the one or more tracked objects, and a location vector for each of the one or more tracked objects.

4. The method of claim 3, wherein each of the one or more tracked objects are classified in accordance with a function of the size vector, the velocity vector, the location vector, and the level of deformation of a corresponding tracked object.

5. The method of claim 4, further comprising the step of adjusting a class of a given tracked object in accordance with at least one of a time of day, an environment, and an observation that the given tracked object belongs to a specific class.

6. The method of claim 1, wherein the step of computing a level of deformation further comprises the steps of:

computing a plurality of histograms of oriented gradients for a given tracked object in the plurality of frames;

detecting at least one change in the plurality of histograms of oriented gradients over the plurality of frames for the given tracked object; and

calculating a level of deformation for the given tracked object using a weighted function, such that a variable of the weighted function is the at least one detected change in the plurality of histograms of oriented gradients over the plurality of frames.

7. The method of claim 6, wherein the weighted function accounts for one or more periods of non-movement of the given tracked object.

8. The method of claim 6, wherein the level of deformation of the given tracked object is compared to one or more object deformation models.

9. The method of claim 8, wherein the one or more object deformation models define a distribution of level of deformation for one or more object classes.

10. An apparatus for classifying one or more objects in at least one video, wherein the at least one video comprises a plurality of frames, the apparatus comprising:

a memory; and

at least one processor coupled to the memory and operative to;

track one or more objects in the plurality of frames;

compute a level of deformation for each of the one or more tracked objects in accordance with at least one change in a plurality of histograms of oriented gradients for a corresponding tracked object; and

classify each of the one or more tracked objects in accordance with the computed level of deformation.

11. The apparatus of claim 10, wherein a given tracked object is classified as at least one of a person and a vehicle.

12. The apparatus of claim 10, wherein the at least one processor is further operative to compute at least one of a size vector for each of the one or more tracked objects, a velocity vector for each of the one or more tracked objects, and a location vector for each of the one or more tracked objects.

13. The apparatus of claim 12, wherein each of the one or more tracked objects are classified in accordance with a function of the size vector, the velocity vector, the location vector, and the level of deformation of a corresponding tracked object.

14. The apparatus of claim 13, wherein the at least one processor is further operative to adjust a class of a given tracked object in accordance with at least one of a time of day, an environment, and an observation that the given tracked object belongs to a specific class.

15. The apparatus of claim 10, wherein the at least one processor is further operative to compute a level of deformation by:

computing a plurality of histograms of oriented gradients for a given tracked object in the plurality of frames;

detecting at least one change in the plurality of histograms of oriented gradients over the plurality of frames for the given tracked object; and

calculating a level of deformation for the given tracked object using a weighted function, such that a variable of the weighted function is the at least one detected change in the plurality of histograms of oriented gradients over the plurality of frames.

16. The apparatus of claim 15, wherein the weighted function accounts for one or more periods of non-movement of the given tracked object.

17. The apparatus of claim 15, wherein the level of deformation of the given tracked object is compared to one or more object deformation models.

18. The apparatus of claim 17, wherein the one or more object deformation models define a distribution of level of deformation for one or more object classes.

19. A system for classifying one or more objects in at least one video, the system comprising:

an input component for obtaining the at least one video, such that the at least one video comprises a plurality of frames;

a feature extractor for tracking one or more objects in the plurality of frames and computing a level of deformation for each of the one or more tracked objects in accordance with at least one change in a plurality of histograms of oriented gradients for a corresponding tracked object;

an object classifier for classifying each of the one or more tracked objects in accordance with the computed level of deformation;

an adaptation component for adjusting a class of a given tracked object in accordance with at least one of a time of day, an environment, and an observation that the given tracked object belongs to a specific class; and

a calibration tool for calibrating a classification of the one or more tracked objects by at least one of partitioning at least one view associated with the at least one video into at least one region, such that the one or more tracked objects are classified in accordance with their location in reference to the at least one region, and defining at least one normalized size of the one or more tracked objects within at least one view associated with the at least one video in accordance with one or more object models, such that the one or more tracked objects are classified in accordance with the at least one defined normalized size.

20. A computer program product for classifying one or more objects in at least one video, wherein the at least one video comprises a plurality of frames, the computer program product comprising:

a computer-readable medium;

a first program instruction to track one or more objects in the plurality of frames;

a second program instruction to compute a level of deformation for each of the one or more tracked objects in accordance with at least one change in a plurality of histograms of oriented gradients for a corresponding tracked object; and

a third program instruction to classify each of the one or more tracked objects in accordance with the computed level of deformation,

wherein the first, second, and third program instructions are stored on the computer-readable medium.