基于JND的压缩感知图像编码  

Image Encoding of Compressive Sensing Based on JND

作　　者：宋允东[1] 王永芳[1,2] 商习武[1] 张兆杨[1,2] 

机构地区：[1]上海大学通信与信息工程学院,上海200072 [2]新型显示技术及应用集成教育部重点实验室,上海200072

出　　处：《电视技术》2012年第14期15-18,共4页Video Engineering

基　　金：国家自然科学基金项目(60972137);国家自然科学基金重点项目(60832003)

摘　　要：传统的数据采集都要遵循奈奎斯特采样定理的两倍以上带宽取样,而压缩感知理论突破原有采样定理的限制,为从少量数据重建原始数据提供了可能性。提出了基于JND压缩感知的稳健性图像编码方法,该新方法将压缩感知理论应用于图像编码,并引入JND模型来提高信号的稀疏性。实验结果表明,提出的方法大大降低了压缩感知的重建时间,同时也提高了图像的重建质量。The traditional data acquisition is under Nyquist sampling theorem, that is,samphng rate must De at least twice mc signal bandwidth. However, compressive sensing broke through the original sampling theorem, whict, provides a possibility for original data recovery from a small amount of data. A method of compressive sensing image robust encoding based on JND model is proposed. Compressive sensing theorem is applied to image coding combined JND model, which improves the sparsity of image signal. Experimental results demonstrate that the proposed method reduces recovery time of compressive sensing greatly, and improves the image quality.

关 键 词：压缩感知 稳健性 JND模型 稀疏性 

分 类 号：TN391.41[电子电信—物理电子学]