盲源分离图像加密的相关运算解密法.docVIP

盲源分离图像加密的相关运算解密法 林秋华, 党杰, 殷福亮 (大连理工大学 电子与信息工程学院，辽宁 大连 116023 ) 摘 要：在盲源分离图像加密方法中，盲源分离解密的多幅图像在顺序和灰度值上存在不确定性，且解密强相关图像前需要去相关预处理。为了解决这两方面问题，利用密钥图像已知且与明文图像不相关的特点，提出了一种基于相关运算的图像解密新方法。该方法根据加密方程计算加密图像与密钥图像的相关阵；通过忽略较小的相关阵，得到与明文图像几乎相同的解密图像。仿真实验和性能分析结果表明，该方法解密质量好、速度快，且适于各种图像。 关键词：盲源分离；；； 中图分类号：TN911.72 文献标识码：Correlation calculation decrypting for BSS-based image cryptosystem LIN Qiu-hua, DANG Jie, YIN Fu-liang (School of Electronic and Information Engineering, Dalian University of Technology, Dalian 116023, China) Abstract: For a blind source separation (BSS) based method for encrypting multiple images, the multiple decrypted images by BSS may have changed order and reversed pixel values. In addition, decorrelation preprocessing must be performed prior to decryption of strongly correlated images. To solve the two problems, a novel method for image decryption was proposed based on correlation calculation since the key images are known for legal users and are independent of the original images. Specifically, the correlation matrix between the encrypted images and the key images was calculated according to the encryption equation, the decrypted images are then obtained by ignoring the smaller correlation matrix between the original images and the key images. The results of computer simulations and performance analyses demonstrate that the proposed method can provide much higher quality and much faster speed compared to the BSS decryption, and also can directly decrypt varied images. Key words: blind source separation; image encryption; image decryption; correlation calculation 1 引言 由于公共互联网和数字多媒体技术的快速发展，资源共享达到了前所未有的程度。与此同时，对于有些数字产品，拥有者希望它们能够安全地传输给合法用户，而不能被非法用户所获得。数字图像作为一种网络上传输的重要信息，其安全传输与存储问题受到了越来越广泛的。一方面，人们对隐私日益重视，不希望私人图片为他人所知；另一方面，在某些领域，如远程医疗系统，法律上要求医学照片必须安全地传输到患者手中。此外，在军事领域，对图像的安全保密性要求更高。图像加密则是保障图像数据安全的重要手段之一。 盲源分离(BSSlind source separation)是近十几年来信号处理领域的热门课题。它在源信号和混合均未知的情况下，能够从混合信号中分离得到源信号估计。利用源信号间统计独立性进行分离独立分量分析ICA, independent component analysis)。目前，盲源分离已被应用在很多领域，如