Name
Abstract | ||
|---|---|---|
The task of visual tracking is to deal with dynamic image sequence. Traditional object representation in tracking algorithms using the image-as-vector subspace learning are easy to result in the problem of the curse of dimensionality and the loss of local structural information from the original image. In this paper, we present a novel online object tracking algorithm by using 2DLPP (Two-Dimensional Local Preserving Projections) manifold learning model. The proposed 2DLPP algorithm adopts a low dimensional eigenspace representation to reflect appearance changes of the target. It can preserve local structural information and directly extract features from image matrices, thereby the method facilitates the tracking task. Furthermore, the new method can update the feature basis recursively, and the computation becomes more efficient for online manifold learning of dynamic object. Finally, we apply the 2DLPP method to visual tracking in the particle filter framework. Experiment results demonstrate the effectiveness of the proposed method in different image sequences where the object undergoes large pose, scale, and lighting changes. |
Year | Venue | Keywords |
|---|---|---|
2014 | Fusion | particle filtering (numerical methods),image representation,visual tracking,2dlpp tracking,learning (artificial intelligence),online object tracking algorithm,low dimensional eigenspace representation,scale change,nonlinear changes,dimensionality curse,feature extraction,object tracking,image sequences,manifold learning,dynamic image sequence,particle filter framework,image-as-vector subspace learning,pose change,local structural information loss,two-dimensional local preserving projection manifold learning model,object representation,appearance model,lighting change,image matrices,2dlpp manifold learning,lighting,learning artificial intelligence,particle filters,visualization |
DocType | Citations | PageRank |
Conference | 1 | 0.36 |
References | Authors | |
0 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Huanlong Zhang | 1 | 1 | 0.36 |
| shiqiang hu | 2 | 1 | 1.04 |
| Lingkun Luo | 3 | 1 | 0.70 |
| Xiaolu Ke | 4 | 1 | 0.70 |