Adaptive Bayesian compressed sensing based on sub-block image

In this paper, a novel algorithm for image sampling and reconstruction is proposed based on Bayesian compressed sensing and sub-block image. Under our proposed scheme, firstly, the image of interest is divided into sub-blocks for reducing recovery time of the image. Secondly, every sub-block across the image is sampled adaptively with diverse sampling rate via compressed sensing skill in the term of each sub-block´s energy. Lastly, a number of sub-blocks are recovered adaptively by using the prior information of neighboring sub-block recovered already. Comparing with the traditional compressed sensing method, our proposed method can recover the image accurately with fewer measurements and less time consumption. Experimental results show the validity and practicality of our proposed method obviously.