Lossless image compression based on a fuzzy linear prediction with context based entropy coding

A novel method for reversible compression of 2D and 3D data is presented. It consists of a spatial prediction followed by context-based classification and arithmetic coding of the outcome residuals. Prediction of a pixel to be encoded is obtained from the fuzzy-switching of a set of linear predictors, i.e., yielding linear combinations of surrounding pixels. The coefficients of each predictor are calculated to minimize prediction MSE for pixels belonging to a cluster in the hyperspace of graylevel patterns lying on a preset causal neighborhood. In the 3D case, pixels both on the current slice and on previously encoded slices may be used. The size and shape of the causal neighborhood, as well as the number of predictors to be switched, may be chosen before running the algorithm to determine the trade-off between coding performances and computational cost. The method exhibits impressive performances, for both 2D and 3D data, mainly thanks to the optimality of predictors, due to their skill in fitting patterns of data.