Optimal Contours Detection By Using A Directional Filtering

P Gouton and I Foucherot (France)

Keywords

Directional filtering, Ridge-line detector, grain boundary extraction, wavelet filter

Abstract

Contour detection remains one of the major points in the segmentation of images. Several surveys have been carried out on the subject. Indeed, the techniques of shape recognition in artificial vision, the search for defects in industrial parts or the characterisation of certain illnesses in medical imaging are based on the extraction of image contours. In this article, we are presenting an optimal filter for the detection of ridge-line. An extension in directional wavelet is proposed in order to take into account all the main directions of the image (or the multiscale character of the required information).

Important Links:



Go Back