Securely Compressive Sensing Using Double Random Phase Encoding

Abstract

Listen

Translate article

Recently, the weakness of existing compressive sensing process from the perspective of the chosen-plaintext attack is discovered. Some algorithms directly use Gaussian matrix as the measurement matrix to do linear dimension reduction projection, which will fail to resist chosen-plaintext attack. To enhance the security and performance of compressive sensing process, double random phase encoding based block compressive sensing is designed, which is chaos-based random phase encoding in fractional Fourier domain for each image block. Moreover, image encryption method using DRPE-based block compressive sensing-combined random phase encoding is proposed. The experimental results demonstrate that the proposed encryption method not only achieves high security level but also has better reconstruction quality compared with other existing encryption methods. Keywords: Securely compressive sensing, fractional Fourier transform, random phase encoding