Proof of Two Kinds of Fingerprint Feature Extraction CNN

The researches and applications of image processing based on the cellular neural network (CNN) have made great progress. The fingerprint feature extraction CNN are two kinds of CNN, which are able to extract the endings and bifurcations, two important features in a fingerprint image. This paper makes a further contribution to this topic. We first present the global tasks and the local rules of these two kinds of CNN. Then, we proof the correctness of those local rules mathematically.