Nonlinear wavelets and BP neural networks Adaptive Lifting Scheme

The Lifting Scheme provides us a flexible and easy way for constructing wavelet, which enables us to construct wavelet according to application needs. Its flexibility allows us to introduce nonlinearity into wavelet and modify it based on signal processed. The Adaptive Lifting Scheme provided the way of adjusting the filters via updater U and predictor P in lifting stages according to signal characters. But the lackness of self-learning ability of existing Adaptive Lifting Scheme is a big shortcoming. In this paper, BP neural networks is introduced into lifting scheme. It is used to replace the updater U and predictor P respectively. Experiment shows that BP neural networks work well in lifting stages.