image processing - Why don't we use motion vector data from video in optical flow? -


all implementation saw in optical flow in opencv uses video array of frames , implement optical flow on each image. involves slicing image nxn block , searching velocity vector.

although motion vector in video codec misleading , not contain motion information, why don't use check block has motion , run optical flow on blocks ? shouldn't fasten process ?

opencv universal image processing framework. takes in frames, not compressed video, algorithms.

you can write video decoder hands out info displacement codec opencv – codec-specific , isn't in scope of opencv itself.


Comments

Popular posts from this blog

sql - Time in datatime field -

javascript - How to split the success output from an Ajax post? -

java - sharing object across different classes -