A Bidirectional Cross-Scrambling Medical Image Encryption Scheme Incorporates Compressed Sensing and Its Application in IoMT
Qiang Lai, Lina Ji
Abstract
Medical images exhibit a wide range of gray scales and are frequently affected by noise and artifacts, necessitating high security and robustness during transmission. A novel medical image encryption scheme combined with compressed sensing is proposed. This scheme utilizes the reconstructed hyperchaotic map as tool for generating random sequences. Through multi-faceted dynamic behavior analysis of the map, it has been verified that it possesses excellent unpredictability and an ultra-wide key space, as well as its hardware implementation lays a foundation for practical image encryption. By decomposing the image into bit-planes, cross scrambling and ascending diffusion are applied to different bit-planes. Combined with compressed sensing, the medical image is transformed into ciphers of varying sizes to achieve the purpose of secure transmission. This approach significantly enhances encryption efficiency by reducing image size. Comprehensive multi-dimensional performance tests demonstrate its superior effectiveness in medical image encryption, providing strong security performance for the application of the scheme in the Internet of Medical Things (IoMT).