On efficient quantization for image recompression

It is well known that both the rate and the distortion of re-compressed images depend primarily on the ratio between the new and the old quantization steps used. In this paper we provide a theoretical basis to this observation, and introduce an efficient algorithm to select the recompression quantization step. Our approach is based on the structure of the quantizer and the distribution of DCT coefficients in subband coding. These results can be readily generalized to re-quantization of other data types. Our conclusion is that the proposed approach could be instrumental in recompression of still images, while offering straightforward generalization to higher and lower dimension signals.