Rotationally invariant wavelet shrinkage

Abstract. Most two-dimensional methods for wavelet shrinkage are ef-ficient for edge-preserving image denoising, but they suffer from poor ro-tation invariance. We address this problem by designing novel shrinkage rules that are derived from rotationally invariant nonlinear diffusion fil-ters. The resulting Haar wavelet shrinkage methods are computationally inexpensive and they offer substantially improved rotation invariance.