A hybrid DCT-VQ based approach for efficient compression of color images

In this paper, we propose a novel color image compression scheme to obtain better reconstructed image as well as high compression ratio. The scheme is mainly based on Vector Quantization, which is one of the well known lossy image compression techniques. During VQ process, we design a common codebook for three color components R, G and B, and to train an effective codebook and speed up the VQ process, we include DCT into VQ operation. This effectively increases PSNR value, reduces blocking artifact during image reconstruction and enhances the overall processing time. The proposed VQ scheme is tested on a number of real color images decomposed into different blocks sizes. The performance has been rated based on different objective measures. The experimental results show that the proposed hybrid scheme is superior to the conventional vector quantization scheme.