期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2015
卷号：8
期号：11
页码：111-118
DOI：10.14257/ijsip.2015.8.11.11
出版社：SERSC
摘要：A blind separation algorithm was proposed for the type of permuted alias images with morphological diversity in this paper. Firstly, permuted alias image mode in sparse domain was presented and separation scheme based on sparse domain was proposed. Non-subsampled Contourlet transform and Waveatom transform are respectively used as two type of dictionaries for piecewise cartoon and texture image. Then the permuting image can be separated from permuted alias image by morphological component analysis algorithm. The results show that our algorithm can separate effectively texture image from the permuted alias image not being affected by size, location number and types of texture image for a permuted alias image composed by piecewise smooth part and texture part
