Medical Image Compressed Sensing Based on Contourlet

Perceived from the definition of compressed sensing (CS), the sparser the signal is, the better the recovery will be. Meanwhile, the third-generation wavelet-contourlet is able to sparsely represent signals and detect the singularity of smooth curve. Taking into account the mixed noise from random projection of CS model, we are trying to carry out the following: let the image transformed into contourlet domain, followed by random observation and carried out using basis pursuit (BP) optimization. The recovery result is obtained after the threshold denoising and inverse contourlet transformation. The experiment results show that the idea is feasible. Compared with other algorithms, the quality of reconstruction is higher.