Tracking demonstration

The user delineates approximately the initial locations in a selected frame and specifies the depth ordering of the objects to be tracked. A first motion-based segmentation is then obtained through an initial application of a Region Growing algorithm. The partition map is sequentially tracked from frame to frame using motion compensation and location prediction. Translational motion is assumed for the moving objects, and local intensity or colour average might be used as additional feature. A post-processing procedure regularizes the object boundaries over time.

The original sequences are the following MPEG-4 test sequences:

The extracted foreground layers are given in the following sequences:

Segmentation and tracking result on the Interview sequence

For further information please see the following publication:
I. Grinias and G. Tziritas, Motion segmentation and tracking using a seeded region growing method,
European Signal Processing Conference, 1998.

Last Modification: November 8, 2000