A Novel Approach on Color Extended Visual Cryptography for General Access Structures using Error Diffusion

Abstract: This thesis proposes a novel approach on color extended visual cryptography for General access structures using the principle of Error diffusion. Conventional visual secret sharing schemes suffers from a management problem, so that dealers cannot identify each share visually. This problem can be solved by the extended visual cryptography scheme (EVCS) by adding a meaningful cover image in each share. The previous approaches involving the EVCS for GAS suffers from a pixel expansion problem and also needs a sophisticated codebook design for various encryption schemes. This paper proposes a general approach to solve the above mentioned problems. This approach can be used for color secret images in non computer aided decryption environments. Color Extended VC encrypts a color secret message or image into color halftone shares. This paper proposes a novel encryption algorithm for the encryption of color secret message using the concept of Jarvis Error diffusion halftoning using Jarvis matrix permutation and Inverse Halftoning using Neural network based method. In visual secret sharing based on Halftone VC, the continuous halftone image is first transformed in to a halftone image and then encrypted using extended VSS. These halftone shares are then error diffused concurrently to give visually pleasing effect. The proposed approach consists of two phases. The first phase of the algorithm based on a given access structure, constructs a set of pixel-expansion-free shares using Jarvis Error diffusion Halftoning. The second phase of the algorithm adds a cover image on each share directly via stamping algorithm based on LSB replacement thereby removing the pixel expansion entirely. In the stamping algorithm, Secret image can be hiding itself in the cover image. Finally secret image will be perfectly reconstructed by inverse halftoning algorithm using neural network method. Secret image can be reconstructed by stacking the qualified set. Comparisons with previous approaches show the superior performance of this newly proposed method.

Keywords: EVCS;GAS;LSB;HT;VSS