Lossless image coding based on lifting wavelet using discrete-time cellular neural network with multi-templates

The lifting wavelet is a well-known method for lossless image compression, and it provides an entirely spatial domain interpolation of the transform, as opposed to the traditional frequency domain based constructions. In this paper, we propose a new lossless image coding technique based on lifting wavelet using discrete-time cellular neural network (DT-CNN) with multi-templates. The advantage of our proposal method is that the output function of the DT-CNN is exploited to consider the nonlinear quantization error which is not considered in the conventional lifting method using linear filters. Additionally, our method improves the prediction for the edge region by using multi-templates of the DT-CNN. The simulation results show a better coding performance compared with the conventional method.