Perceptual compressive sensing based on contrast sensitivity function: Can we avoid non-visible redundancies acquisition?

Conventional compressive sensing (CS) attempts to acquire the most important part of a signal directly. In fact, CS avoids acquisition of existed statistical redundancies of a signal. Since the sensitivity of the human eye is different for each frequency, in addition to statistical redundancies, there exist perceptual redundancies in an image which human eye could not detect them. In this paper, we propose a novel CS approach in which the acquisition of non-visible information is also avoided. Hence, we could expect a better compression performance. We deploy the weighted CS idea to consider these perceptual redundancies in our model. Moreover, the block-based compressed sensing is favorable since it has some advantages: (a) It needs low memory to store the sensing matrix and sparsifying basis. (b) All blocks can be reconstructed in parallel. Therefore, we apply our proposed scheme in the block-based framework to make it practical to use. Simulation results verify the superiority of our proposed method compared to the other state-of-the-art methods.