Deformable models are physics-based models that deform under the laws of Newton mechanics, in particular, by the theory of elasticity expressed in the Lagrange dynamics. These models (snakes, balloons, etc.)
are active models that represent a powerfull tool to treat physical objects in images. The goal of the course is to show deformable models as an appropriate technique for different problems of Computer Vision: segmentation,
3D reconstruction, non-rigid motion recovering, etc.
Special attention will be paid to show different applications of the deformable model technique.
Part I, Introduction on snakes
Part II, Fundamentals of snakes
Part IIb, B-snakes
Part III, High-dimensional snakes
Part IV, Variations of snakes and their application to the segmentation problem, segmentation of radiographs and facial images
Part V, Region-based and color-based snakes
Part VI, Topologically invariant (geodesic) snakes
Part VII, Statistic snakes
Part VIII, Statistic snakes applied to facial analysis
Bibliography on deformable models,
Extended bibliography on deformable models.
The snake's tale began from:
Kass, M. and Witkin, A. and Terzopolous, D., "Snakes: Active Contour Models", in International Conference on Computer Vision, London, pp. 259-268, 1987.
For any suggestions and new discoveries on snakes, please, send me an e-mail!