Image denoising with true pixel presence detection procedure
M. Sruthi, L. Mredhula, M. A. Dorairangaswamy · 2014
This paper proposes the benefit of applying Bayesian approach to the wavelet coefficients for denoising of images. One of the main disadvantages of simple thresholding is the blurring of the image and the problem of low quality. This is minimized by modifying the thresholding function in an adaptive manner. The result will be enhanced if an a priori distribution is assumed to the wavelet coefficients and then finding an efficient estimation method. While comparing with the results of adaptive thresholding a better peak signal to noise ratio is observed. Also the presence of the signal is detected in the intermediate step before estimation process. The simple decision method in the signal estimation, Maximum Likelihood Decision Criterion is applied in this work.