Hands-On Genetic Algorithms with Python

Structural Similarity (SSIM)

The SSIM index was created to be used to predict the image quality that's produced by a given compression algorithm by comparing the compressed image to the original one. Rather than calculating an absolute error value, which is done by the MSE method, for example, SSIM is perception-based and considers changes in structural information, as well as effects such as brightness and texture in the images.

The metrics module of the scikit-image library provides us with a function that calculates the structural similarity index between two images. When both images are represented using the OpenCV (cv2) library, this function can be used directly, as follows:

SSIM = structural_similarity(cv2Image1, cv2Image2)

The value ...