Linear Estimation and Wavelet Approach to Image Denoising

A linear estimator termed as time domain constrained (TDC) estimator combined with wavelet filtering, is applied for restoration of images corrupted by additive white Gaussian noise (AWGN). Since, signal distortion and residual noise cannot be minimized simultaneously, the TDC estimator keeps the residual noise energy below a threshold while minimizing the signal distortion. Further reduction in the residual noise is achieved by a wavelet-based denoising process involves threshold shrinkage of corrupted wavelet coefficients. The proposed approach is tested with different images corrupted by AWGN at different signal-to-noise values and compared with the conventional wavelet and Wiener filtering