Title :
Tracking video objects using active contours
Author :
Castaud, M. ; Barlaud, M. ; Aubert, G.
Author_Institution :
CNRS UNSA, Sophia Antipolis, France
Abstract :
We address the problem of tracking video objects through several frames. We present a new tracking algorithm using the active contour theory framework. Our first contribution is to define a new tracking criterion combining geometrical and region-based features. Our second contribution is to provide the differentiation of the criterion and to deduce the evolution equation of the active contour. Finally, we apply this general tracking framework to statistical region-based and geometrical descriptors.
Keywords :
differentiation; minimisation; object detection; optical tracking; statistical analysis; video signal processing; active contour theory; differentiation; evolution equation; geometrical features; minimization; region-based features; video object tracking; Active contours; Deformable models; Differential equations; Image segmentation; Image sequence analysis; Motion estimation; Shape; Tracking; Video sequences;
Conference_Titel :
Motion and Video Computing, 2002. Proceedings. Workshop on
Print_ISBN :
0-7695-1860-5
DOI :
10.1109/MOTION.2002.1182219