for every video, there are about 250 frames, but it's metadata is split into about 130 time instances, need a way to uniformly map the frames into the time values without biases to one side. np.array_split() makes it biased towards the start with a lower number of frames per time instance in the last few frames. Need to fix this imbalance.
for every video, there are about 250 frames, but it's metadata is split into about 130 time instances, need a way to uniformly map the frames into the time values without biases to one side.
np.array_split()
makes it biased towards the start with a lower number of frames per time instance in the last few frames. Need to fix this imbalance.