i need help . my program is for repositioning automation. has to compare two image, ref and image. im using many algorithms to check similarity , and then im trying to do a shift between the images(shifting one based on the other) to get higher similarity and less error. i need to get the optimal shift , that will give me in all the algorithms im comparing with (SSIM, pixel by pixel, MSE, cross correlation, residuals) lets say in one i got first time similarity 7%, i need to find the shift that will raise the similarity to 98% - im talking about very sensitive way of shifting(sensitive for small differences) i tried phase correlation with iterations, it didnt give me what i want. i tried minimizing MSE between the two images with (delta x delta y) . i still dont get the 5 algorithms to 98% (actually none of them till now) Continue reading...
