I would compute the gradient of a video sequence. I have $x$, $y$, $t$ dimensions so I would obtain $dx$, $dy$, $dt$ referred to the all video sequence. I know that I have to use a kernel filter to calculate the gradient by $x$ and $y$, but I have doubt how to compute the $dt$... I read about finite differences, how can I apply it to my sequence? I would also know, if it is correct to consider the sum of the $dx$ (i.e) of all frames, to have the total gradient of the video sequence (for the x dimension, and doing the same to calculate the y video gradient). Is it?
thank you
SOLUTION:
Even if nobody was able to give me good answers (maybe for my bad formulation of the question) I find a solution to my problem that a want to share with you.
1) x-derivatives and y-derivatives can be computed by shifting Sobel filter in every frames of the video sequence. Than we can save every frame result.
2) t-derivatives (time gradient) can be computed by finite differences By following the notation:
t-derivative= frame at t = i+1 - frame at t=i where i=1,2,3,...,FrameSequenceNumber.
Thank you.