OpenCV 3 Computer Vision with Python Cookbook

How it works...

To find a solution to the orthogonal Procrustes problem, we applied SVD to the multiplication product of two matrices: the matrix composed of initial points and another one composed of points after rotation. The rows in each matrix are (x, y) coordinates of corresponding points. The SVD approach is well-known and gives stability to noise outcomes. cv2.SVDecomp is a function that implements SVD in OpenCV. It accepts a matrix (MxN) to decompose and returns three matrices. The first returned matrix is a rectangular diagonal matrix of size MxN, with positive numbers on the diagonal called singular values. The second and third matrices are a left-singular vector matrix and a conjugated transpose of a right-singular vector matrix, ...