Medical image compression using vector quantization and system error compression

A novel medical image compression scheme based on vector quantization (VQ) is proposed in this paper. The advantages of the technique are not only that it yields high compression ratio but also that it maintains a peak signal-to-noise ratio (PSNR). This new method involves three steps. First, we present a codebook design using discrete wavelet transform (DWT), fuzzy C-means (FCM), and support vector machine (SVM) algorithms. Second, we improve the bit rate using the Huffman coding theme as a method of eliminating the redundant index. Finally, we supplement the system with error compensation to improve the PSNR. With the proposed method, we are able to achieve a bit rate improvement of 24.00% and a PSNR of 10.96% over the conventional method.