Grayscale Image Compression Using DCT Transformation

In this experiment, we will employ Discrete Cosine Transform (DCT) to compress grayscale images. The implementation involves dividing the image into 8x8 pixel blocks and applying 2D-DCT transformation to each block using matrix operations. We will calculate compression ratios by comparing original file sizes with compressed versions after quantization and zig-zag scanning. The process includes implementing threshold-based coefficient retention where higher frequency coefficients are discarded based on compression requirements. We will compare reconstructed images with different compression ratios by applying inverse DCT transformation and evaluating quality metrics like PSNR (Peak Signal-to-Noise Ratio). Through this experiment, we can gain deep insights into DCT transformation principles and its applications in image compression. This helps us better understand image processing and compression algorithm workflows, while enabling systematic evaluation and comparison of visual effects at different compression levels through quantitative analysis and visual inspection.