Edge-adaptive super-resolution image reconstruction using a Markov chain Monte Carlo approach
Jing Tian, Kai‐Kuang Ma · 2007
In our recent work, the Markov chain Monte Carlo (MCMC) technique has been successfully exploited for performing super-resolution image reconstruction. Despite its powerful performance, it usually suffers from that the reconstructed highresolution image is too smooth to lose much detail information. To further enhance the edge and detail information in the reconstructed high-resolution image, an edge-adaptive MCMC super-resolution approach is proposed in this paper. Steered by an edge map of the desired high-resolution image, the proposed method can adaptively enhance the edge information at the edge pixel positions while exploiting the conventional MCMC SR at the rest pixel positions. Experimental results are presented to demonstrate the superior performance of the proposed method.