Hi All,
Is there anyway to do batch processing with the OpenCV Background Subtraction algorithms such as MOG2, etc…? As in, I want to pass in several images as a tensor and then process them in a single function call. I am hoping it would improve the processing time for a large number of simultaneous videos (similar to deep learning approaches in batches).
no, there isn’t (what do you want to achieve with this idea ?).
((all bg subtraticon algos need continuous input, a sequence, all you’d get from feeding in a batch of images is some “fast-forward”))
I think you have misunderstood the question. I am not asking to pass in images from the same video (which would not make sense). I am asking if I have say 100 independent videos which are synchronized, and I want to run background subtraction on them independently, can this be parallelized through GPU/CUDA? For example, a single operation on a single video frame may take 10ms to execute, however, I am hoping that there is a batch operation that can improve the processing time, as 10ms for each video from a set of 100 videos is obviously not desirable.