Object tracking using momentum and acceleration vectors in a motion estimation system
There is provided a method and apparatus for motion estimation in a sequence of video images. The method comprises a) subdividing each field or frame of a sequence of video images into a plurality of blocks, b) assigning to each block in each video field or frame a respective set of candidate motion vectors, c) determining for each block in a current video field or frame, which of its respective candidate motion vectors produces a best match to a block in a previous video field or frame, d) forming a motion vector field for the current video field or frame using the thus determined best match vectors for each block, and e) forming a further motion vector field by storing a candidate motion vector derived from the best match vector at a block location offset by a distance derived from the candidate motion vector. Finally, steps a) to e) are repeated for a video field or frame following the current video field or frame. The set of candidate motion vectors assigned at step b) to a block in the following video field or frame includes the candidates stored at that block location at step e) during the current video field or frame The method enables a block or tile based motion estimator to improve its accuracy by introducing true motion vector candidates derived from the physical behaviour of real world objects.
1. A method for motion estimation in a sequence of video images, comprising:
selecting a best match motion vector for a block in a current video field or frame from a set of candidate motion vectors based on a best match of a candidate motion vector to a block in a previous video field or frame;
determining a position in memory corresponding to a block in a next video field or frame by applying the best match vector to the block in the current video field or frame;
storing a prediction candidate vector equal to the best match vector at the determined position in memory; and
using the stored prediction candidate vector to perform motion estimation in said next video field or frame.
2. The method of claim 1 , further comprising storing multiple candidate motion vectors at a block location based on a priority system.
3. The method of claim 2 , wherein said priority system comprises:
storing a metric representing the quality of the best match in said previous video field or frame; and
selecting candidate motion vectors for storing according to the stored quality metric.
4. A method for motion estimation in a sequence of video images, comprising:
selecting a best match motion vector for a block in a current video field or frame from a set of candidate motion vectors based on a best match of a candidate motion vector to a block in a previous video field or frame;
determining a first position in memory corresponding to a block in a next video field or frame by applying the best match vector to the block in the current video field or frame;
storing a prediction candidate vector equal to the best match vector at the determined first position in memory;
selecting a best match motion vector for a block in the next video field or frame from a set of candidate vectors including said prediction candidate vector based on a best match of a candidate vector to a block in said current video field or frame;
calculating a change in motion vector corresponding to a difference between a best match vector in said next video field or frame and a stored prediction candidate vector for a block in said next video field or frame;
adding said change in motion vector to said best match motion vector for said block in said next video field or frame to produce an acceleration candidate vector;
determining a second position in memory corresponding to a block in a second next video field or frame by applying said acceleration candidate vector to said block in said next video field or frame;
storing said acceleration candidate vector at the determined second position in memory; and
using the stored acceleration candidate vector to perform motion estimation in said second next video field or frame.
5. Apparatus for motion estimation in a sequence of video images, comprising:
a selector configured to select a best match motion vector for a block in a current video field or frame from a set of candidate motion vectors based on a best match of a candidate motion vector to a block in a previous video field or frame;
a memory configured to store a prediction candidate vector equal to the best match vector at a position in memory corresponding to a block in a next video field or frame determined by applying the best match vector to the block in the current video field or frame; and
a motion estimator configured to use the stored prediction candidate vector to perform motion estimation in said next video field or frame.