Statistical modeling and treshold selection of wavelet coefficients in lossy image coder

An algorithm of wavelet domain data quantization to improve compression efficiency is presented. Conditional probability model of adjacent tin scale-spatial sense) magnitudes was applied as better approximation of wavelet coefficient dependencies than marginal data distributions. This model was utilised in threshold data selection proposed as more effective uniform quantization modification than dead-zone increasing. The same conditional model was used in quantization and encoding of quantized magnitudes. Additionally, to fit adaptively threshold value to local image features, estimation of significance expectation was included into thresholding procedure. As a result, more effective low-cost quantization scheme was constructed. It allows significantly increase compression efficiency of images. Experimental Rate-Distortion curve shows the same distortion for decreased bit rates even up to 20% in comparison to standard uniform quantization.