Modifications of uniform quantization applied in wavelet coder
Artur Przelaskowski · 2002
An algorithm of wavelet domain data quantization aimed at improving compression efficiency is presented. Threshold data selection is proposed as a more effective uniform quantization modification than zero-zone increasing. To fit adaptively the threshold value to local image features, the estimation of the significance expectation for each wavelet coefficient was included in the thresholding procedure. The remaining data are uniformly quantized without any changes of bin boundaries. As a result, a more effective low-cost quantization scheme was constructed. It allows us to significantly increase the compression efficiency of images. An experimental rate-distortion curve shows the same distortion for decreased bit rates even up to 20% in comparison to standard uniform quantization. Such a quantization technique was applied in a wavelet coder with optimised schemes of decomposition and zerotree based coding. Its compression efficiency is competitive with the most efficient methods across all natural images tested.