Abstract: Based on the active contour model, a robust algorithm is proposed for extracting and
optimizing valley-ridge lines from point-sampled geometry (PSG). First, local implicit surface of the
PSG is reconstructed, and the curvature of every sampling point is computed. Potential valley-ridge
points are then obtained by solving the extreme of the main curvature, and potential valley-ridge points
is connected into valley-ridge polylines according to the main direction of the feature points. Finally,
snake model is used to smooth the valley-ridge polyline. Experimental results show that the algorithm is
robust and can generate smooth valley-ridge lines.

Key words: point-sampled geometry, feature extraction, valley, ridge, Snake model