Visual Cryptography Scheme Using Pi-sigma Neural Networks

A visual cryptography scheme (VCS) for general access structures is a method to encode a secret image into n shares for n participants. Only qualified subsets of participants can recover the secret image. The yet known schemes omitted the deference between primary information and secondary one in secret image, and its best known upper bound on the share size is exponential. Besides its information communication rate is high. This paper proposes a new VCS for general access structures using pi-sigma neural networks (VCSPSN). A pi-sigma neural network (PSN) is a class of two-layer feedforward net-work. Compared to other high-order networks, a PSN has a highly regular structure, needs a much smaller number of weights and less training time. Using PSN´s capability of large-scale parallel classification, VCSPSN reduces the information communication rate greatly, makes best known upper bound polynomial, and distinguishes the deferent information in secret image. This paper firstly introduces the structure of PSN, and then constructs the VCSPSN and analysis its security and capability.