A Study of Secret Image Sharing Schemes

• Main Authors: Ching-Fen Lee, 李青芬
• Other Authors: Justie Su-Tzu Juan
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/73027276099498502428

碩士 === 國立暨南國際大學 === 資訊工程學系 === 97 === Secret image sharing scheme can efficiently solve the problems that are digit materials, military documents and important images could be forged, falsified and stolen, etc. Up to now, there are many scholars to design various secret image sharing schemes. However, most of the schemes have many drawbacks and limits or still can be improved. For example: the computation is too complex; each participant holds too many shares; those schemes only can be applied in specified access structure; cannot share multi-secret images, achieve the capabilities of the multi-use, verification and detection at the same time. Based on the above defects, the goal of this thesis presents secret image sharing schemes with more integrity and multi-function. Therefore, we will divide thesis into three parts: In first part, Shyu and Chen proposed multiple secret images sharing schemes in 2008, but the secret images could be found by a set of unqualified participants in their general access structure sharing scheme. Therefore, we propose two multi-secret images sharing schemes (t, n)-MSISS1 and GMSISS1. The sizes of the public image and time complexity of these two schemes are better than Shyu and Chen’s schemes. Moreover, any set of the unqualified participants cannot found the secret images in these two schemes. In second part, Shi et al. proposed a (t, n)-threshold secret sharing scheme for image encryption in 2008, but there are some drawbacks in this scheme: the phenomenon shapes of the secret image could be found by using the public image, the width must be equal to the height, and the size of the secret images must be the same upon sharing multi-secret images. Therefore, we propose two multi-secret images sharing schemes (t, n)-MSISS2 and GMSISS2 to solve these problems. Furthermore, the proposed (t, n)-MSISS2 and GMSISS2 can respectively achieve sharing multi-secret images, multi-use, and each participant only holds one share, and GMSISS2 can be used for any general access structure. In third part, Zhao et al. proposed image secret sharing scheme to identify cheaters in 2009. But these exists the problem of disguise, and cannot achieve multi-use forever in Zhao et al.’s scheme. Therefore, we propose a verifiable and detectable multi-secret images sharing scheme with general access structure VDGMSISS to provide a verifiable procedure to authenticate identity of valid participants and a detectable method to check accuracy of the share. In addition, the proposed VDGMSISS can achieve some properties, such as sharing multi-secret images, multi-use, without security channel and be used for any access structure.