Systems and methods for object tracking and localization in videos with adaptive image representation
A system and method are provided for object tracking in a sequence of images of a scene acquired by an imaging device. The method includes generating a representation of an object and its surroundings; generating a tracking score based on the representation of the object and its surroundings, and a classification scheme; detecting a peak in the tracking score corresponding to a location of the object in the image; and adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object.
1. A method for object tracking in a sequence of images of a scene acquired by an imaging device, the method comprising:
generating a representation of an object and its surroundings;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image;
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object; and
performing an iterative adaptive adjustment to update image representations and classifiers.
2. The method according to claim 1 further comprising performing a multiple resolution and multiple scale procedure to encompass changes of the object size and scale in a consecutive frame.
3. The method according to claim 1 further comprising using a high dimensional representation of the object and its surroundings, wherein the high dimensional representation is a numeric representation of image data for the image.
4. The method according to claim 3 further comprising using an automatically learned image representation.
5. The method according to claim 1 further comprising using a set of weak classifiers to generate the tracking score for a new observation.
6. The method according to claim 1 further comprising performing an optimization procedure to maximize the tracking score for the object of interest.
7. The method according to claim 1 further comprising performing an iterative adjustment based on strong observations by rejection of the outliers.
8. The method according to claim 1 further comprising generating a strong classifier by combining weak classifiers.
9. A non-transitory computer readable medium comprising computer executable instructions for object tracking in a sequence of images of a scene acquired by an imaging device, comprising instructions for:
generating a representation of an object and its surroundings;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image;
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object; and
performing an iterative adaptive adjustment to update image representations and classifiers.
10. The non-transitory computer readable medium according to claim 9 further comprising performing a multiple resolution and multiple scale procedure to encompass changes of the object size and scale in a consecutive frame.
11. The non-transitory computer readable medium according to claim 9 further comprising using a high dimensional representation of the object and its surroundings, wherein the high dimensional representation is a numeric representation of image data for the image.
12. The non-transitory computer readable medium according to claim 11 further comprising using an automatically learned image representation.
13. The non-transitory computer readable medium according to claim 9 further comprising using a set of weak classifiers to generate the tracking score for a new observation.
14. The non-transitory computer readable medium according to claim 9 further comprising performing an optimization procedure to maximize the tracking score for the object of interest.
15. The non-transitory computer readable medium according to claim 9 further comprising performing an iterative adjustment based on strong observations by rejection of the outliers.
16. The non-transitory computer readable medium according to claim 9 further comprising generating a strong classifier by combining weak classifiers.
17. A system comprising a processor and memory, the memory comprising computer executable instructions for object tracking in a sequence of images of a scene acquired by an imaging device, comprising instructions for:
generating a representation of an object and its surroundings;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image;
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object; and
performing an iterative adaptive adjustment to update image representations and classifiers.
18. A method for object tracking in a sequence of images of a scene acquired by an imaging device, the method comprising:
generating a high dimensional representation of an object and its surroundings, wherein the high dimensional representation is a numeric representation of image data for the image;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image; and
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object.
19. The method according to claim 18 further comprising performing a multiple resolution and multiple scale procedure to encompass changes of the object size and scale in a consecutive frame.
20. The method according to claim 18 further comprising using a set of weak classifiers to generate the tracking score for a new observation.
21. The method according to claim 18 further comprising using an automatically learned image representation.
22. The method according to claim 18 further comprising performing an adaptive adjustment in the learned representative representation of the image in the following image frames.
23. The method according to claim 18 further comprising performing an adaptive adjustment in the set of classifiers in order to compute the tracking score.
24. The method according to claim 18 further comprising performing an iterative adaptive adjustment to simultaneously update image representations and classifiers.
25. The method according to claim 24 further comprising performing an optimization procedure to maximize the tracking score for the object of interest.
26. The method according to claim 24 further comprising performing an iterative adjustment based on strong observations by rejection of the outliers.
27. The method according to claim 18 further comprising generating a strong classifier by combining weak classifiers.
28. A non-transitory computer readable medium comprising computer executable instructions for object tracking in a sequence of images of a scene acquired by an imaging device, comprising instructions for:
generating a high dimensional representation of an object and its surroundings, wherein the high dimensional representation is a numeric representation of image data for the image;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image; and
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object.
29. The non-transitory computer readable medium according to claim 28 further comprising performing a multiple resolution and multiple scale procedure to encompass changes of the object size and scale in a consecutive frame.
30. The non-transitory computer readable medium according to claim 28 further comprising using a set of weak classifiers to generate the tracking score for a new observation.
31. The non-transitory computer readable medium according to claim 28 further comprising using an automatically learned image representation.
32. The non-transitory computer readable medium according to claim 28 further comprising performing an adaptive adjustment in the learned representative representation of the image in the following image frames.
33. The non-transitory computer readable medium according to claim 28 further comprising performing an adaptive adjustment in the set of classifiers in order to compute the tracking score.
34. The non-transitory computer readable medium according to claim 28 further comprising performing an iterative adaptive adjustment to simultaneously update image representations and classifiers.
35. The non-transitory computer readable medium according to claim 34 further comprising performing an optimization procedure to maximize the tracking score for the object of interest.
36. The non-transitory computer readable medium according to claim 34 further comprising performing an iterative adjustment based on strong observations by rejection of the outliers.
37. The non-transitory computer readable medium according to claim 28 further comprising generating a strong classifier by combining weak classifiers.
38. A system comprising a processor and memory, the memory comprising computer executable instructions for object tracking in a sequence of images of a scene acquired by an imaging device, comprising instructions for:
generating a high dimensional representation of an object and its surroundings, wherein the high dimensional representation is a numeric representation of image data for the image;
generating a tracking score based on the representation of the object and its surroundings, and a classification scheme;
detecting a peak in the tracking score corresponding to a location of the object in the image; and
adjusting the representation of the object and its surroundings and the classification scheme according to one or more new appearances of the object.