Image Coding Using Wavelet-based Compressive Sampling

In this paper, we proposed a novel coding scheme is proposed using wavelet-based CS framework for nature image. First, two-dimension discrete wavelet transform (DWT) is applied to a nature image for sparse representation. After multi-scale DWT, the low-frequency sub-band and high-frequency sub-bands are re-sampled separately. According to the statistical dependences among DWT coefficients, we allocate different measurements to low-and high-frequency component. Then, the measurements samples can be quantized. The quantize samples are entropy coded and forward correct coding (FEC). Finally, the compressed streams are transmitted. At the decoder, one can simply reconstruct the image via l(1) minimization. Experimental results show that the proposed wavelet-based CS scheme achieves better compression performance against the relevant existing solutions.