Hexagonal Image Compressionusing Singular Value Decomposition in Python

With the advent of multimedia technologies in last two decades, there is a widespread need for efficient storage and transmission of data. Dealing with the vast information interchange in this digital era, image compression for reduction in byte size of graphics image file without loss of image quality to an acceptable level becomes the large interest area. Inspired from the biological models of human fovea, hexagonal image processing has gained a lot of attention in artificial intelligence era that deals with the application of image processing system that combines the benefits of biologically motivated structures. In this paper a singular value decomposition (SVD) over hexagonal image compression which is a missing stone in computer vision which provides higher packing density, higher angular symmetry and uniform connectivity. Due to lack of developments in hexagonal imaging devices, different resampling methods like alternate pixel shift method, half pixel shift method, pseudo hexagonal pixel method for sourcing hexagonal images. SVD is one of the powerful cutting-edge technology for image compression algorithms. SVD based image compression is performed on hexagonal grid and is compared with square grid using different parameters like compression ratio, compression size, PSNR and MSE using PYTHON's SVD function. SVD based hexagonal image achieves the goal of compression by preserving good image quality at higher compression ratios, high computational efficiency, provides low mean square error (MSE), acceptable compression size depending on application and high peak to signal ratio (PSNR).