基于压缩感知的加密图像多描述编码方法

Encryption Image Multiple Description Coding Method Based on Compressive Sensing

基于压缩感知和多描述编码理论,提出了一种鲁棒性好且易于实现的加密图像压缩方法.通过对稀疏变换后的图像进行交织抽取,对所得到的各系数块分别进行加密运算,然后经过分块压缩测量、量化和打包形成多描述压缩码流.解码端根据接收到的各码流,通过求解优化问题和解密运算重构原图像.实验结果表明,相对于同类方法,该方法能有效地提高图像传输的鲁棒性以及重构图像的主观和客观质量.

A novel encryption image compression method was proposed based on compressive sensing and multiple description coding theory in this paper. This method is robust to packet loss and bit error,and has the advantage of easy implementation. The original image is firstly performed scarifying transforma- tion,followed by encryption operation for each coefficients block generated by interleaving extraction in transform domain. The encrypted image compression is achieved by block compressive sensing and multiple descriptions coding. At the decoder side, the original image is recovered by solving an optimization problem and decryption operation. Experiment results show that compared with the similar method, the proposed method can effectively improve the robustness of the image transmission and enhance the reconstructed image quality in both PSNR and visual.