Image Super-Resolution by Structural Sparse Coding
Jie Ren, Jiaying Liu, Mengyan Wang and Zongming Guo
Super resolution is a very ill-posed inverse problem and strong prior model about the high resolution image should be incorporated to regularize the SR problem. Recently, a emerging technique which relies on sparse and redundant representations for image patches has attracted a lot of attentions. In these patch-based methods, neighbor patches' relationship via overlapped regions is only to keep smoothness of reconstructed high-resolution image, so the prior is not always strong enough to regularize super resolution when the observed low-resolution image loses structure information. In this paper, we propose to improve the performance of the sparsity-based method by incorporating the structural correlations between neighboring patches. Experimental results demonstrate that the pro- posed algorithm outperforms the sparsity-based base- lines in both objective and subjective quality.
© 2012 WangMengyan