Title :
Image filtering based on piecewise linear models
Author_Institution :
D.E.E.I., Trieste Univ.
Abstract :
Piecewise linear (PWL) models are very attractive for image processing due to their simplicity and effectiveness. A new filtering architecture adopting parameterized PWL functions is presented. The proposed approach takes into different account the correlation between the central pixel of a moving window and its neighbors in two subwindows. As a result, the filtering errors can be reduced and a very accurate restoration of the image data can be achieved. The sensitivity to a change of the parameter settings is also investigated
Keywords :
Gaussian noise; correlation methods; filtering theory; image restoration; piecewise linear techniques; filtering error; image data restoration; image filtering; image processing; piecewise linear model; Filtering; Finite impulse response filter; Gaussian noise; Image denoising; Image processing; Image restoration; Noise cancellation; Nonlinear filters; Piecewise linear techniques; Pixel;
Conference_Titel :
Imaging Systems and Techniques, 2004. (IST). 2004 IEEE International Workshop on
Conference_Location :
Stresa
Print_ISBN :
0-7803-8591-8
DOI :
10.1109/IST.2004.1397271
