A Method to Improve the Performance of JPEG Algorithm

JPEG is one of the most widely used methods for image compression. In the JPEG image compression method, the quantization table used in the quantization step plays an important role. The quantization table significantly determines the quality of the image and the amount of compression. Therefore, the creation of the quantization table is seen as an optimization problem. In this article, rate-distortion approach, human visual system approach and meta-heuristic approaches have been explained in order to create more effective quantization tables. Then, a new table was obtained by taking the average of the neighbors of that value instead of each value in the JPEG quantization table. The results were tested on 5 different images using different quality factors. It has been found that the proposed table increases the compression ratio by 3.52% compared to the standard JPEG table, almost without sacrificing the quality of the images.