Wyner-Ziv coding of image using compressed sensing

Compressed sensing (CS) is a technique for acquiring and reconstructing a signal utilizing the prior knowledge that it is sparse or compressible. In this paper, we present a Wyner-Ziv image coding framework, which combines the traditional image coding and compressed sensing. In our method, the image separate into two parts according to pixel location: the down-sampling image is encoded by traditional intra-frame coding and the rest part is encoded through CS method. In order to obtain a better recover performance, the whole image is jointly reconstructed using the two part data based on image sparse representation. The experimental results are reported to demonstrate the effectiveness of the proposed coding method.