Image compression using adaptive lifting scheme based on minimum mean square error criterion

摘要

Wavelet transform has demonstrated excellent compression performance with natural images. The framework of lifting offers the flexibility for developing adaptive wavelet transforms. In this paper, adaptive prediction and update techniques based on the local properties of the images are proposed. A non-linear wavelet transform that chooses the number of vanishing and preserving moments in the prediction and update lifting steps adaptively according to the underlying local signal characteristics is presented. The spatial adaptivity criterion is based on the optimum interpolating function and Minimum Mean Square Error (MMSE). This adaptive lifting transform appears promising for image compression.