Why does warpAffine take a dSize parameter?

asked 2014-01-29 18:41:10 -0500

xaffeine gravatar image

It seems to me that the combination of the input matrix size and the coefficients in the transform matrix dictate the output size. The document doesn't say how the dsize parameter is used. Is it for cropping?

I've had trouble getting my code to work right. Maybe I should look at the source code of cvGetQuadrangleSubPix, but what I really want to do is extract a patch of pixels bounded by a RoundedRect (with non-zero rotation, obviously). I am trying to implement

void ExtractPatch( const cv::Mat & srcImg, cv::Mat & destImg, const cv::RotatedRect & rr );
edit retag flag offensive close merge delete