Research of image sparse algorithm based on compressed sensing

The sparse image representation plays an important role in the image processing. Improved layer discrete cosine transform (DCT) and Contourlet transform are two sparse algorithms. This paper compared compression and recovery results of the two sparse algorithms based on compressed sensing, while they were used to process the same image. The results show that, compared with the improved layer DCT in the compressed sensing image application, the compressed sensing image sparse algorithm based on Contourlet transform can sparsify image signal and keep more details of the image. For the same number of measurements, the average peak signal to noise ratio (PSNR) is improved about 4 dB.