Morphological Representation of DCT Data for Image Coding

Recent success in discrete cosine transform (DCT) image coding is mainly attributed to recognition of the importance of data organization and representation. Currently, there are several competitive DCT-based coders such as Xiong et al.’s DCT-based embedded image coding (EZDCT), Davis and Chawla’s significance tree quantization (STQ), and Zhao et al.’s embedded zerotree image coder based on hierarchical DCT (EZHDCT). In the wavelet context, morphological representation of wavelet data has achieved the best compression performance for still image coding. The representatives are Servetto et al.’s morphological representation of wavelet data (MRWD) and Chai et al.’s significancelinked connected component analysis (SLCCA). In this paper, we first point out that the block-based DCT by proper reorganization and representation of its coefficients can have the similar characteristics to wavelet transform, such as energy compaction, crosssubband similarity, decay of magnitude across subband, etc. This finding will widen DCT applications relevant to image compression, image retrieving, image recognition and so on. We then present a novel image coder utilizing these characteristics by morphological representation of DCT coefficients (MRDCT). The experiments show that MRDCT is among the state-of-art DCT-based image coders reported in the literature. For example, for the Lena image at 0.25 bpp, MRDCT outperforms JPEG, STQ, EZDCT and EZHDCT by 1.0 dB, 1.0 dB, 0.3 dB and 0.1 dB in PSNR, respectively. This outstanding performance is achieved without using any optimal bit allocation procedure. Thus both the encoding and decoding procedure are fast. PERFORMANCE COMPARISION (PSNR [dB]) ON LENA IMAGE