視覺機密分享(Visual Secret Sharing, VSS)是一種針對影像進行加密的技術,其最大特色為在解密時不需要任何計算,而是透過人類的視覺系統(Human Visual System)。此外,視覺機密分享可以視為針對影像的機密分享(Secret Sharing),因此視覺機密分享繼承了機密分享的優點。其概念是將一張機密影像加密成兩張無意義的分享影像(Share),我們唯有疊合這兩張分享影像才能重建機密影像。隨機網格(Random Grid)是實現視覺機密分享的一種方法。近年來,在此領域中有越來越多研究是針對如何將多張機密影像加密成兩張分享影像,這一類的研究被稱作多重機密影像(Multiple Secret-Images)分享。 本篇論文提出基於平移(Shifting) 隨機網格的三種品質改良機制。張於2012年所提出的,利用平移隨機網格之多重視覺機密影像配置系統,理論上能加密無限多張。但實際上其失真率(Distortion) 會隨機密影像的數量增加而顯著提升,因此能加密的機密影像數量就會受到限制。在本篇論文中,我們提出三種改良的方法,其中兩種之概念是增加分享影像的像素,來降低失真率; 第三種是平均打散分享影像上的黑色像素來增進影像品質(Image Quality)。這三種方法的影像品質都會勝過原先的方法,因此我們就能加密更多張機密影像。
Visual secret sharing (VSS) is a technique of image encryption, and we decrypt secret images by human visual system without any computations. Furthermore, we can assume visual secret sharing as visual version of secret sharing, so visual secret sharing inherits all advantages of secret sharing. The concept of VSS is encrypting a secret image into two meaningless shares. We can only decrypt the secret image by stacking these shares together. Random grid is an approach to implement visual secret sharing. In recent years, many research for encrypting many secret images into two shares are proposed. The topic of these research is multiple secret-images sharing. This thesis proposes three quality-enhanced schemes basing on shifting random grids. In theory, we can encrypt unlimited number of secret images by using the AOSRG scheme which be held by Chang in 2012. However, the distortion of AOSRG scheme is large when the number of secret images 6. Because of this reason, the number of secret image to be encrypted is restricted. In this thesis, we propose three improved schemes. Two of those are adding the number of pixels on shares to decrease the distortion, and the other is evenly distributing black pixels on shares to increase the image quality. The image quality of those schemes are better than AOSRG scheme. Therefore, we can encrypt more secret images by using the proposed schemes.