LRJPEG: A Luminance Reduction based Modification for JPEG Algorithm to Improve Medical Image Compression

Abstract

The size of multimedia contents is one of the biggest challenges, especially the storage and transmission the same during emergency situations. Compression is one of the fundamental techniques to address this problem. In this work, we have developed an improved Joint Photographic Experts Group (JPEG) compression algorithm called LRJPEG (Luminance Reduction JPEG) using optimized modifications in the luminance quantization table for both color and grayscale images. The quantization in an important step in the JPEG compression process. The default quantization table ( $8\times 8$ ) depends on both the chrominance and luminance components of an image. We have modified the $8\times 8$ default luminance quantization table. To assess the performance of the improved JPEG algorithm, we have used different color and grayscale medical images during the compression process. We have evaluated CR (Compression Ratio), PSNR (Peak signal-to-noise ratio), and MSE (mean squared error) of both standard and a modified quantization table. The results show that the performance of JPEG compression using modified luminance quantization table has improved compared with the default luminance quantization table.