1.
Supriya Tiwari
– Reader (et&t Dept) Rcet Bhilai, Chhattisgarh, India
2.
Nitin Naiyar
– Reader (et&t Dept) Rcet Bhilai, Chhattisgarh, India
Abstract
This paper proposes a comparison between adaptive wavelet thresholding, Non-Parametric Regression and Hybrid method based on SNR (Signal to Noise Ratio) and MSE (Mean Square Error). First method uses Wavelet Transform for de-noising of image. The proposed threshold is closed-form and simple, because it depends on data-driven estimates of the parameters so it is adaptive to each sub band. In second method, an image De-Noising procedure is proposed, based on Non-parametric estimation of a jump surface from noisy data. The third method is a hybrid of the above mentioned two methods. In this the output of adaptive wavelet thresholding method is fed as input to the Non-parametric regression method.
