Image Compression Using Optimal Quantization

An image compression algorithm based on quantization of the quasi-two-dimensional spectrum quantization using the Walsh transform is proposed. This paper develops an image compression algorithm that differs from similar methods by using a quasi-two-dimensional spectrum, which simplifies the transition to frequency space as much as possible. The quasi-two-dimensional spectrum undergoes decimation, followed by uniform or optimal quantization of remaining elements. Resulting matrix is compressed by Deflate algorithm. Application of this method allowed to compress images 5–8 times with acceptable restored images quality. Another distinctive feature of the algorithm is the use of the entire image instead of a block structure, which allows you to avoid artifacts of such processing methods. This algorithm can be used in image transmission systems over communication channels. This paper develops a new algorithm of image compression that has a compression rate up to 0.6 bit per pixel with PSNR 31