Lifting-Based Lossless Parallel Image Coding on Discrete-Time Cellular Neural Networks

Abstract

Although the nonlinear interpolative dynamics of discrete-time cellular neural network (DT-CNN) is an effective method for prediction-based image coding schemes such like the lifting wavelet, the iterations of CNN dynamics are a bottleneck of processing time. This paper presents a novel lossless parallel image coding method based on lifting scheme using DT-CNNs. In the proposed method, split steps of the lifting scheme are extended in order to achieve fast image compression by parallel processing, and the subsampled image is interpolated by using the nonlinear interpolative dynamics of DT-CNN. Since the output function of DT-CNN works as a multi-level quantization function, the proposed method composes the integer lifting scheme for lossless coding. The experimental results show that the processing cost is greatly reduced by the proposed coding scheme