Joint thresholding and quantizer selection for decoder-compatible baseline JPEG

摘要: This paper introduces a novel, image-adaptive, encoding scheme for the baseline JPEG standard. In particular, coefficient thresholding, quantization matrix (Q-matrix) optimization, and adaptive Huffman entropy-coding are jointly performed to maximize coded still-image quality within constraints of syntax. Adaptive coding has been addressed in earlier works: by Ramchandran Vetterli (see IEEE Trans. on Image Processing, Special Issue Compression, vol.3, p.700-704, September 1994), where fast rate-distortion (R-D) optimal thresholding was described, Wu Gersho Proc. Inter. Conf. Acoustics, Speech Signal vol.5, p.389-392, April 1993) Hung Meng (1991), R-D optimized Q-matrix selection performed. By formulating an algorithm which optimizes these two operations jointly, we have obtained performance comparable more complex, "state-of-the-art" schemes: "Lenna" image at 1 bpp, our achieved PSNR 39.6 dB. result represents gain 1.7 dB over with customized entropy coder, even slightly exceeds published Shapiro's vol.41, p.3445-3462, December wavelet-based scheme. Furthermore, choice appropriate visually-based error metrics, noticeable subjective improvement as well.