SURE and LAWMLShrink Image Denoising Methods with MATLAB Implementation

The SURE (Stein's Unbiased Risk Estimate) and LAWMLShrink (Locally Adaptive Window-based Maximum Likelihood Shrinkage) methods represent advanced image denoising approaches implementable through MATLAB. These algorithms employ wavelet transform decomposition combined with adaptive thresholding strategies to effectively suppress various noise types while preserving image edges and textures. The MATLAB implementation typically involves key functions including wavelet decomposition (wavedec2), threshold calculation using SURE optimization, and localized noise estimation through LAWMLShrink's window-based processing. The core algorithm workflow comprises: 1) Multi-level wavelet decomposition of noisy images 2) Adaptive threshold computation using SURE risk minimization 3) Local variance estimation via sliding window operations 4) Inverse wavelet reconstruction (waverec2) with optimized coefficients. This implementation provides configurable parameters for wavelet types (Daubechies, Symlets), thresholding rules (soft/hard), and window sizes for localized processing. The code includes performance metrics such as PSNR calculation and visual comparison tools to evaluate denoising effectiveness across different noise levels. This MATLAB implementation serves as a practical reference for researchers and developers working on image restoration projects, offering a balanced approach between computational efficiency and denoising quality. The provided documentation covers both theoretical foundations and practical implementation details for reliable integration into image processing pipelines.