An Image Compression Scheme Using AC Predictions

An improved image compression scheme using AC coefficient prediction of Hadamard transform is presented in this paper. In the prediction phase, information of the nearest neighbor dc values of Hadamard transform blocks is utilized to predict the center block AC coefficients. Linear programming based optimization technique is used to predict the AC coefficients. Results indicate that this method is superior in terms of improving peak signal to noise ratio (PSNR) and minimizing blocking artifacts .