Abstract | ||
---|---|---|
An eigenshape kernel based mean shift tracker is proposed in this paper. In contrast with the symmetric constant kernel used in the traditional mean shift tracker, this tracker employs eigenshape to construct an arbitrarily shaped kernel that is adaptive to object shape. Therefore, background information is adaptively excluded from the target. Furthermore, the eigenshape kernels are integrated with color and gradient features, which enhance tracking robustness. Experiments demonstrate that this tracker outperforms the traditional mean shift tracker significantly especially when target shape deformation, target occlusion and background clutter occur. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1109/ICCVW.2011.6130468 | ICCV Workshops |
Keywords | Field | DocType |
target occlusion,mean shift tracker,target shape deformation,tracking robustness,symmetric constant kernel,background clutter,feature extraction,object tracking,image sequences,color features,arbitrarily shaped kernel,object shape,eigenvalues and eigenfunctions,gradient features,eigenshape kernel based mean shift,human tracking,image colour analysis,kernel,histograms,mean shift,shape | Kernel (linear algebra),Computer vision,Histogram,Pattern recognition,Clutter,Computer science,Robustness (computer science),Feature extraction,Video tracking,Artificial intelligence,Mean-shift | Conference |
Volume | Issue | ISBN |
null | null | 978-1-4673-0062-9 |
Citations | PageRank | References |
0 | 0.34 | 5 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Chun-Mei Liu | 1 | 245 | 41.30 |
Changbo Hu | 2 | 613 | 34.71 |
J. K. Aggarwal | 3 | 428 | 50.43 |