Memory-efficient video tracking in real-time using direction vectors

1. A computer-implemented method comprising:

receiving, by a processing engine, a sequence of image data;

transforming objects in each frame of the sequence of the image data into direction vectors;

clustering the direction vectors based at least in part on features of the objects;

mapping the direction vectors for the objects in each frame into a position-orientation data structure; and

performing tracking using the mapped direction vectors in the position-orientation data structure.

2. The computer-implemented method of claim 1 further comprising maintaining the direction vectors for each frame of the sequence of image data.

3. The computer-implemented method of claim 1 , wherein clustering the direction vectors is based at least in part on smallest inter-frame distances and inter-frame orientation differences between the direction vectors of adjacent frames.

4. The computer-implemented method of claim 1 , wherein the position-orientation data structure comprises a plurality of bins that are arranged according to orientation information and position information.

5. The computer-implemented method of claim 4 , wherein a range corresponding to an orientation for each bin and position for each bin are configurable parameters.

6. The computer-implemented method of claim 1 further comprising storing the clustered direction vectors for each frame in the same location in the position-orientation data structure.

7. The computer-implemented method of claim 1 , wherein the features of the objects include position information and orientation information.

8. The computer-implemented method of claim 1 further comprising:

comparing corresponding direction vectors for a current frame with direction vectors of a previous frame;

tracking the object based at least in part on the comparison; and

re-projecting proposed targets for the object based on tracking the object.

9. The computer-implemented method of claim 1 further comprising:

identifying a missing object in a current frame;

performing a look-up operation in the position-orientation data structure based on the feature of the object of the current frame;

retrieving missing object data from the position-orientation data structure; and

displaying output data including the retrieved missing object data.

10. The computer-implemented method of claim 9 , wherein the lookup operation is performed on a region of the position-orientation data structure that is less than all regions of the position-orientation data structure.

11. A system for performing object tracking, the system comprising:

one or more processors; and

at least one memory, the memory including instructions that, upon execution by at least one of the one or more processors, cause the system to perform a method for performing memory-efficient tracking, the method comprising:

receiving a sequence of image data;

transforming objects in each frame of the sequence of the image data into direction vectors;

clustering the direction vectors based at least in part on features of the objects;

mapping the direction vectors for the objects in each frame into a position-orientation data structure; and

performing tracking using the mapped direction vectors in the position-orientation data structure.

12. The system of claim 11 further comprising storing the clustered direction vectors for each frame in the same location in the position-orientation data structure, wherein clustering the direction vectors is based at least in part on smallest inter-frame distances and inter-frame orientation differences between the direction vectors of adjacent frames.

13. The system of claim 11 , wherein the position-orientation data structure comprises a plurality of bins that are arranged according to orientation information and position information.

14. The system of claim 11 , wherein the features of the objects include position information and orientation information.

15. The system of claim 11 further comprising:

comparing corresponding direction vectors for a current frame with direction vectors of a previous frame;

tracking the object based at least in part on the comparison; and

re-projecting proposed targets for the object based on tracking the object.

16. The system of claim 11 further comprising:

identifying a missing object in a current frame;

performing a look-up operation in the position-orientation data structure based on the feature of the object of the current frame, wherein the lookup operation is performed on a region of the position-orientation data structure that is less than all regions of the position-orientation data structure;

retrieving missing object data from the position-orientation data structure; and

displaying output data including the retrieved missing object data.

17. A computer program product for performing object tracking, the computer program product comprising a computer-readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

receive a sequence of image data;

transform objects in each frame of the sequence of the image data into direction vectors;

cluster the direction vectors based at least in part on features of the objects, wherein the features of the objects include position information and orientation information;

map the direction vectors for the objects in each frame into a position-orientation data structure, wherein the position-orientation data structure comprises a plurality of bins that are arranged according to orientation information and position information; and

perform tracking using the mapped direction vectors in the position-orientation data structure.

18. The computer program product for performing 17 , wherein:

clustering the direction vectors is based at least in part on smallest inter-frame distances and inter-frame orientation differences between the direction vectors of adjacent frames; and

the program instructions are further executable by the processor to cause the processor to store the clustered direction vectors for each frame in the same location in the position-orientation data structure.

19. The computer program product for performing 17 , wherein the program instructions are further executable by the processor to cause the processor to:

compare corresponding direction vectors for a current frame with direction vectors of a previous frame;

track the object based at least in part on the comparison; and

re-project proposed targets for the object based on tracking the object.

20. The computer program product for performing 17 , wherein the program instructions are further executable by the processor to cause the processor:

identify a missing object in a current frame;

perform a look-up operation in the position-orientation data structure based on the feature of the object of the current frame, wherein the lookup operation is performed on a region of the position-orientation data structure that is less than all regions of the position-orientation data structure;

retrieve missing object data from the position-orientation data structure; and

display output data including the retrieved missing object data.