可变可视密码

Abstract

传统的可视密码在加密时会产生像素扩张, 结果使分存图像比秘密图像大许多倍, 尤其是应用在灰度和彩色图像上, 其扩张的倍数更是惊人. 传统的可视密码都是单点加密, 本文在Hou的m点加密的基础上, 提出任意点加密可视密码, 即在加密的时候可以对任意个点进行加密, 我们称之为可变可视密码. 操作的时候, 对秘密图像的r个点同时进行加密, 当r=m时, 该加密就是像素不扩展可视密码; 当r>m时, 该加密得到的就是分存图像缩小的可视密码(r的增大会降低解密图像的对比度); 当r<m时, 该加密得到的就是分存图像扩大的可视密码. 随着r的增大, 分存图像会变小, 同时对比度也会降低. 对r个点同时加密的时候, 需要计算r个点中黑点的个数b, 对于有b个黑点的加密, 在r次加密中, 保证有b次使用黑色像素加密矩阵B_1加密, r-b次使用白色像素加密矩阵B_0加密. 可变可视密码一方面解决了传统可视密码像素扩张的问题; 一方面它非常灵活, 能使分享图像小于、等于或大于加密图像, 从而能有效减少存储空间, 或在存储空间和图像质量之间找到一个平衡点.

Alternate abstract:

Conventional visual cryptography needs to expand pixels and enlarge the size of shares. This situation is more serious for gray-level and color images. Conventional visual cryptography is single-pixel encryption. Based on Hou's m-pixel encryption, this study proposes a visual cryptography of any pixel encryption, which is called variable visual cryptography (VVC). When computing the shares, r pixels of the secret image are encrypted all together. If r=m, the encryption is the visual cryptography without expanding the pixels; if r>m, the encrypted image is smaller than the shared image (the increase of r will reduce the contrast of the decrypted image); if r<m, the encrypted image is larger than the shared image. With the increase of r, the image will be smaller and the contrast will be lower. When encrypting r pixels together, it is necessary to compute the number of black pixels b in r pixels. For encrypting b black pixels in r times, we use black pixel encrypting matrix B_1 for r times , and we use white pixel encrypting matrix B_0 for r-b times. By doing that, it solves the pixel expansion problem of conventional visual cryptography, and it is so flexible that the sharing images can be smaller, equal or larger than the original secret image, so it can reduce the storage space effectively, or find a balance between storage and image quality.