Performance Comparison of Image Block Compressive Sensing Based on Chaotic Sensing Matrix Using Different Basis Matrices

In this paper, a performance comparison of image block compressive sensing (BSC) based in chebyshew chatotic matrix using different sparse basis matrices is studied. These basis matrices are discrete cosine transform (DCT), discrete Hartley transform (DHT), discrete Fourier transform (DFT), discrete wavelet transform (DWT), and discrete Wash Hadamard transform (WHT), respectively. The test image is divided into sub-blocks, which are transformed into sparse domain by sparse basis. Then the conventional recover algorithm OMP is used to evaluate the image quality. With the number of sub blocks is increasing, the quality of the recovered image becomes deterioration in terms to the peak signal-tonoise ratio. In addition, block compressive sensing based on DWT is the best when the number of blocks of image is 1, 2. On the other hand, when the number of sub block is 4 and 8 cases, the quality of the recovered image based on DeT is the best. For all BSe schemes based on sparse expect DWT basis, the number of sub block has very litter effect on the PRNR of the reconstructed images.