An efficient prediction algorithm for image vector quantization.

In this paper, we propose a novel efficient fuzzy prediction algorithm for image vector quantization by using the fact that the pixel values of a block is highly related to those of its adjacent blocks in the same image frame because of its spatial correlation. The experimental results show that our method performs better than other VQ algorithms, such as SOC-VQ, LCIC-VQ, STC-VQ, and DST-VQ, in terms of computational complexity and coding efficiency.