Adaptive Generalized Prediction for Lifting Schemes
Joel Solé, Philippe Salembier · 2006
The lifting scheme is a useful tool to create different types of wavelet decompositions, including adaptive and nonlinear. Generalized lifting is more flexible and can improve lifting results, but the design of generalized prediction and update steps remains difficult for a given application. A design strategy to optimize the prediction step according to the image statistics is established. The criterion aims to minimize the detail signal coefficients energy, The scheme is used for lossless compression of several classes of images without any book-keeping or side information requirements. Promising results are reported for certain classes of images.