ACTIVE CONTOURS APPROACH TO OBJECT TRACKING IN IMAGE SEQUENCES WITH COMPLEX BACKGROUND

Active contour models (''snakes'') are a powerful tool for deformable object tracking in moving images. But the existing snake models are no t well-adapted for tracking corners and objects on a complex backgroun d. In this paper, we present a novel active contour model, the ''Adjus table Polygons'', which is a set of active segments that can fit any o bject shape (including corners). A new energy based on textural charac teristics of objects is also proposed, in order to resolve conflict si tuations while tracking objects on multiple contour background.