Context based SAR data compression using fuzzy logic

In this paper we propose a new image coding technique based on context modeling of wavelet coefficients using fuzzy logic. The scalar codebook was partitioned into four disjointed subsets. A fuzzy logic was used to predict the current subset using previous subsets and the values of the wavelet coefficients in the context. We designed the fuzzy logic using fuzzy rules and membership functions in order to achieve minimal mean square error. The proposed context-based modeling of wavelet coefficients outperforms the trellis quantization in the rate distortion sense by 0.2 dB in SNR.