An embedded image coder based on hierarchical DCT

Debin Zhao,Wen Gao,Yuki Chan

1999-01-01

Abstract:Since Shapiro's work on embedded zerotree wavelet image coding, there have been variations by other researches. In this paper, we show that the zerotree quantizer originally developed for wavelet compression can be effectively applied to hierarchical DCT(HDCT) with excellent results and eases the rate-control problem. In the proposed scheme, the input image is partitioned into a number of 8×8 blocks and a first level DCT is applied to each of these individually. As DC coefficients of neighboring blocks are highly correlated and particularly pronounced the compression results at low bit rates, HDCT is applied to only DC coefficients organized as 8×8 DCT blocks. All the HDCT coefficients within a DCT block are then rearranged into a sub-band structure in which the zerotree quantizer can be employed. The proposed algorithm yields a fully embedded, low-complexity coder with PSNR's from 0.83 to 1.66 dB better than the baseline JPEG for the standard image "Lena". In order to remove the blocking effects in the reconstructed images, especially at low bit rates, a simple but efficient block effect removal method based on Sobel operator and smoothing constraints is also proposed. Experimental results show the proposed deblocking method works well and enhances decoding for decompressed images both objectively and subjectively.