when i use
warpAffine(InputArray src, OutputArray dst, InputArray M, Size dsize);
imshow("dst",dst);
there is visible area and invisible area in "OutputArray dst" according to "Size dsize".
then i think, if dsize is small, it is expected area in "InputArray src" that only need to be warpped and this will be more efficient.
warpAffine work like this way?
In other words,
cropping area and warping or warping whole image (src) and cropping ROI?
(sorry for my poor english.)