Hierarchical enumerative coding of DCT coefficients
L. Öktem, Ruşen Öktem, Jaakko T. Astola · 2002
We propose the employment of hierarchical enumerative coding, an entropy coding method, for the entropy coding of DCT coefficients in a DCT-based image compression scheme. The following configuration is used for the comparisons: the 8/spl times/8 DCT coefficients are first rearranged to obtain a wavelet-transform-like layout. Then, an improved version of the morphological representation of wavelet data (MRWD) algorithm is applied for scanning the coefficients and encoding the locations of nonzero coefficients. Within this configuration, two alternative entropy coding methods, context-based arithmetic coding and hierarchical enumerative coding, are compared. Experimental results show that the performance of hierarchical enumerative coding is as good as that of arithmetic coding. Considering the little computational complexity of hierarchical enumerative coding, the equal compression performance suggests that hierarchical enumerative coding is a feasible alternative for the entropy coding of DCT coefficients.