I would like to divide each frame of a video by an image that I would like to provide as an additional input. Can ffmpeg perform such image calculations? I could not find any examples.
This would be a kind of background correction. The input image is the background, for example the median of a large number of frames. (often people use a backgroundSUBTRACTION, but I want to use division)
currently I am doing this in openCV and python. I process each frame like this:
(vidMed is the background image)
import cv2
vidIn=cv2.VideoCapture(input.avi)
vidOut=cv2.VideoVriter(output.avi)
while true
ret, frame = vidIn.read()
bgDiv=frame/vidMed
vidOut.write(bgDiv)
I was hoping this may run faster using ffmpeg
yes, blend works!
ffmpeg.exe
-i videoIn.avi
-i imageIn.avi
-filter_complex "[1:0] setsar=sar=1 [1sared];[0:0][1sared]blend=all_mode='divide':repeatlast=1[int];
-out.avi
(imageIn.avi is just one frame)