VQ‐based compressive sensing with high compression quality

Abstract

Natural image reconstruction based on compressive sensing (CS) has shown a promising performance in recent years. However, sometimes the restoration precision is not high enough. A novel CS algorithm using vector quantisation (VQ) error is proposed. First, the original image is compressed by VQ due to its extremely high compression ratio and strong ability to preserve details. Then compute the VQ error matrix and ignore the three least significant bits, which makes the error matrix much sparser. Next, to ensure a uniform distribution of sparsity of blocks, the error matrix is scrambled. Since the huge diversity among blocks has been largely reduced, they can be sensed with the same sensing matrix in space domain. At last, the reconstruction effect of the error matrix decides the total restoration performance. Experimental results have demonstrated the proposed method, at low measurement ratio, performs better in the aspects of perception and peak signal‐to‐noise ratio.

Secure Image Authentication Scheme Using Double Random-Phase Encoding and Compressive Sensing

“…Yu et al [15] theoretically proved that the measurement matrix generated by chaotic map satisfies restricted isometry property (RIP), which provided an alternative construction of the CS measurement matrix. In addition, since the sparseness of the signal was utilized to reconstruct the original signal from the measurements, directly assigning raw images as inputs of CS may lead to a worse restoration and may not achieve the ideal compression effect due to the lower sparsity of raw images [16,17].…”