Title :
Flexible filter neighbourhood designation
Author :
Smith, Stephen M.
Author_Institution :
Defence Res. Agency, Chertsey, UK
Abstract :
This paper describes an approach to low level image noise reduction which uses non-linear comparisons to define which parts of the image are closely related to each individual pixel, i.e. what the local smoothing neighbourhood is. The resulting method achieves good smoothing without degrading edges, lines or corners in the image. It is also investigated for its multi-scale properties, and shown to give very little shift in edge position over multiple smoothing scales
Keywords :
edge detection; feature extraction; noise; smoothing methods; SUSAN filter; flexible filter neighbourhood designation; local smoothing neighbourhood; low level image noise reduction; multi-scale properties; nonlinear comparisons; Brightness; Computer vision; Degradation; Filtering algorithms; Image edge detection; Noise reduction; Nonlinear filters; Pixel; Smoothing methods; Statistics;
Conference_Titel :
Pattern Recognition, 1996., Proceedings of the 13th International Conference on
Conference_Location :
Vienna
Print_ISBN :
0-8186-7282-X
DOI :
10.1109/ICPR.1996.546020