Effect of Different Block Sizes on the Quality of JPEG-like Compressed Image

Discrete Cosine Transform (DCT) is a common method in image compression processing. In this method, the original image is first divided into several blocks for processing, and then the information of each pixel is converted to frequency coefficients, and a certain high frequency component is discarded, thus the image is lossy compressed. In this paper, the effect of block sizes on the quality of decompressed image is studied, so as to find out the optimal block sizes. Firstly, we make a theoretical analysis, and then verify my analysis results through experiment. In the experiment, the block sizes range from 1x1 to 512x512. We use different metrics to evaluate the image quality, including MSE, PSNR and SSIM.