Prediction capabilities of Boolean and stack filters for lossless image compression (Q1289152)

Prediction methods for lossless image compression build models of the image and transmit the parameters of the models together with the prediction errors. The paper advocates the use of the Error Entropy criterion (EE) in the design of Boolean filters, which is a natural choice for this problem. New optimal design procedures for Boolean filters are introduced. Stack and FIR-Boolean hybrid filters are proposed as well. Various forms of these filters are considered: global-optimal, block-optimal, adaptive-size-block-optimal, multiresolution. The mean absolute error criterion and its relation to the EE are also studied. Simulation results are reported. They show that, in general, the EE-optimal Boolean predictors are superior and that the other designs may be more suitable in certain circumstances.