本論文提出一個可以處理相異尺寸的多重秘密影像分享技術，在此技術中又有兩種處理方式，先將不同尺寸的秘密影像做統一的處理，再利用雜湊函數還有亂數種子產生出亂數影像，透過亂數影像和秘密影像做布林函式運算，最後產出分享影像，在處理的過程中都是利用基本運算，所以在效能上有很好的效率，在安全性上也達成了秘密影像分享的效果，最後更能處理更多不同尺寸的影像。相較其他學者所提出的分享方法，本論文針對多樣性及安全性兩方面做了更佳的改進。在安全性方面，本文採取有幾張秘密影像就生成幾張亂數影像，在得到分享影像時，不會因為任意兩張影像做互斥或運算而得到秘密影像之資訊，在多樣性方面，因為在做分享運算前，我們就要針對不同影像尺寸來做處理，透過處理後便能直接進行分享動作，最終還原也可還原回原本尺寸大小，使得秘密影像的多樣性有更大的突破。 A Boolean-based multiple secret image sharing technique shares many secret images among participants using low computational Boolean operations. Previous studies dealt with sharing identical sized secret images. However, secret images may include different sizes for usage requirement. Therefore, this paper presents two schemes, named partial sensitivity different sized symmetric sharing-recovery (PDSR) and full sensitivity different sized symmetric sharing-recovery (FDSR), for generating shared images from secret images, in which shared images can have different sizes. These two schemes preserve a significant property of acquiring the same random image from different sized secret images or from shared images. The difference between these two schemes consists in the quality of recovered images when shared images are suffered from attacks. In PDSR, attacks outside the minimum sized area, which is defined as minimum shared image size between secret images, only fail to recover the corresponding secret image. But, attacks within the minimum sized area fail to recover any secret image. In FDSR, any attack on shared image leads to false recovery of any secret image. The proposed scheme modifies previous symmetric secret sharing algorithm to fit our requirement of different sized secret images. Therefore, these two proposed schemes are also symmetric of using the same function to generate shared images by secret images and to recover secret images by shared images. Experimental results show that these two proposed schemes takes similar computation time to share and to recover different sized secret images. Comparing with related Boolean based schemes, these two proposed schemes preserve the most integrated characteristics on sharing secret images and overcome all drawbacks that other schemes occurred.
