Taylor Sun Flower Optimization-Based Compressive Sensing for Image Compression and Recovery.

This article presents a novel image compression and recovery approach based on compressive sensing using a Taylor-based Sunflower Optimization (Taylor-SFO) algorithm. The method operates in two phases: first, compressing medical images by exploiting self-similarity and a three-dimensional (3D) transform; second, recovering the compressed image using a structural similarity index measure (SSIM)-based collaborative sparsity measure (S-CoSM) combined with the Taylor-SFO algorithm, which integrates Taylor series expansion with the sunflower optimization technique. Experimental evaluation on the BRATS 2018 medical image dataset demonstrates that the proposed approach achieves superior image quality, with peak signal-to-noise ratio (PSNR) values up to 57.57 dB and SSIM values up to 0.9412, outperforming existing methods such as BEMD, hyper-chaotic map encryption, cyclic shift compressive sensing, and pulse coupled neural network (PCNN) based techniques. The study highlights the effectiveness of combining non-local 3D sparsity and local 2D sparsity in the recovery phase and suggests future work involving hyper-chaotic compressive sensing algorithms.