Low-complexity and low-memory entropy coder for image compression

In this paper, a low-complexity and low-memory entropy coder (LLEC) is proposed for image compression. The two key elements in the LLEC are zerotree coding and Golomb–Rice (G–R) codes. Zerotree coding exploits the zerotree structure of transformed coefficients for higher compression efficiency. G–R codes are used to code the remaining coefficients in a variable-length codes/variable-length integer manner resulting in JPEG similar computational complexity. The proposed LLEC does not use any Huffman table, significant/insignificant list, or arithmetic coding, and therefore its memory requirement is minimized with respect to any known image entropy coder. In terms of compression efficiency, the experimental results show that discrete cosine transform (DCT)and discrete wavelet transform (DWT)-based LLEC outperforms baseline JPEG and embedded zerotree wavelet coding (EZW) at the given bit rates, respectively. For example, LLEC outperforms baseline JPEG by an average of 2.2 dB on the Barbara image and is superior to EZW by an average of 0.2 dB on the Lena image. When compared with set partition in hierarchical trees, LLEC is inferior by 0.3 dB, on average, for both Lena and Barbara. In addition, LLEC has other desirable features, such as parallel processing support, region of interest coding, and as a universal entropy coder for DCT and DWT.