視覺密碼是一種依靠人眼解密的一種秘密分享方法，在無法使用電腦解密的情況下，它是一個很好的解決方案。大部分的視覺密碼方法由於像素擴展的關係，所以產生的分享影像大小會比原來的機密影像大上許多倍，尤其是應用在灰階影像上，其擴展的倍數更是驚人。本研究提出一種不需要像素擴展的視覺密碼技術；我們的方法利用多點同時加密的概念，每次取m個點進行加密，在這m個點上出現黑點的機率，則是依黑點在基礎矩陣中出現的比例而訂，因此我們可以達成像素不擴展的目標，並確保重疊影像的視覺效果。針對灰階影像，我們利用半色調技術將灰階影像轉換成只有黑與白兩色的半色調影像，使得原本只能應用在黑白影像上的視覺密碼方法，也可以直接應用在灰階影像上。此外，我們利用提高對比的方法來解決還原影像中對比損失的問題，使得還原的影像有更好的視覺效果。由實驗結果可以證明，在不作像素擴展的前提之下，我們的方法不但可以應用在任意的使用結構上，也能確保還原的影像有很好的視覺效果。
Visual Cryptography is a visual secret sharing method. It provides a solution to decrypt secrets without computers. Most Visual Cryptography needs to expand pixels, so that the shares are much larger than the original secret image. The problem of pixel expansion is more serious when it applied to gray-level images. In this paper, we propose a Visual Cryptography for gray-level images without pixel expansion. The main idea of our method is to encrypt m pixels simultaneously. The probability of black pixels filling in these m pixels is depends on the ratio of black pixels in the basis matrix. Therefore, we don't need to expand pixels and still get good visual effect. Besides, we utilize halftone technique to convert gray-level images into bi-level images, which have black and white pixels only; thus our method can be applied to gray-level images easily. In addition, we stretch the contrast of the secret image to offset the lost contrast caused by stacking shares. The experimental results show that our method can be applied to any access structure without pixel expansion and also performs better on the visual effect than other methods.
