New classified vector quantization of images

In this work, a new coding scheme called predictive classified address vector quantization (PCAVQ) is proposed for image compression. In this scheme, a new two-stage classification method based on the three-level block truncation coding technique is first developed to efficient classify blocks of an image into different classes having similar characteristics. The predictive mean removal VQ technique is applied to reduce the blocking effect in decoded images. On the other hand, a new simplified address VQ method is also developed to increase the compression ratio of the overall system. Simulations using real images show that the proposed PCAVQ scheme has about 1.7 to 4 dB improvement in peak signal to noise ratio (PSNR) over the JPEG baseline system.<>