Content based authentication of Visual Cryptography
Visual Cryptography (VC) is perceived and studied as a perfect combination of secret sharing and digital image processing. The basic idea of VC is to split original secret image into several partitions which are also called shares. VC schemes include basic VC, grayscale VC, colour VC and multi-secret VC etc. Despite the security nature of VC in secret sharing, one of the common problems of current application of VC shares is that it lacks authentication. Previous related researches have proven the possibility of VC cheating through different methods. Attackers are able to complete both cheating and modification on VC process without being noticed by VC participants. Currently available authentication schemes for VC are derived from the view of utilizing additional shares and blind authentication.
This research analyses effective authentication methods using 2D barcodes and embedding binary codes into VC shares for authentication purpose. A scheme of embedding 2D barcodes into VC shares to prevent cheating will be presented to enhance the use of VC in implementation. The embedding process includes four steps: resolution adaption, image matching and replacement, barcode selection and secret recovery. The aim of this research is to propose a method of embedding 2D barcode into VC shares, thereby strengthening the cheat prevention ability of VC shares by applying the security feature of 2D barcode into VC. As an international standard of reading guidance for the blind people, Braille has been widely used as an effective communication channel. In this thesis, we will also explain Braille encoding and explain how it is applied to handle the authentication problem in VC. Similar to the use of 2D barcode in VC, the utilization of Braille in VC is also attributed to the similarity of structure and construction between Braille cells and VC shares.
Even though the research of visual cryptography is based on the combination of image processing and cryptography, knowledge of VC authentication related to digital image processing and cryptography has not been fully utilized in the past years. In this thesis, the analysis of both visual features and cryptographic features of VC will be presented and utilized to assist VC authentication. The visual features of VC in this thesis include moments, histogram, centroid, entropy and Tamura Texture. Compared to those existing methods, the contribution of this research is to propose an authentication scheme of integrate those features with Hash code and digital signature so as to be embedded into VC shares.