Hi guys.. I am using the haar cascades to detect faces, and to make it do it faster I work on grayscale image, and wondering if should change the dimensions of the image it applied to.. I mean it makes totally sense that applying it to a smalle image creates less work.
But the face detection is only used to find create a ROI of the face. So how do i detect a face in small picture, and transfer those parameter into the original picture. such that i will know exactly where the face is in i the "big picture" ?