Efficient variable rate vector quantization using quadtree segmentation

Abstract

In video and image compression, vector quantization can be fixed rate or variable rate depending on how the encoding codebook is structured. Previous variable rate vector quantization schemes using greedy tree growing algorithm partitions images into fixed size blocks and has not exploited the correlation in the low detail blocks. A new and efficient variable rate vector quantization scheme based on variable block size quadtree segmentation is proposed in this paper. Simulation results show that the proposed variable block size variable rate vector quantization scheme achieved better subjective image quality with higher PSNR at a reduced bit rate as compared with the fixed block size variable rate vector quantization scheme.