Robust visual tracking based on Gabor feature and sparse representation

Li, Weiguang; Hou, Yueen; Lou, Huidong; Ye, Guoqiang

doi:10.1109/robio.2012.6491234

Cited by 2 publications

(1 citation statement)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is widely used in intelligent surveillance, human-computer interface, vehicle navigation, and traffic control, etc. Although many tracking algorithms have been presented in the past decades, the tracking of the non-stationary appearance of objects undergoing significant pose, illumination variations and occlusions still remains a challenge for the community [1].In general, a tracking algorithm consists of three components [2]: an appearance model which estimates the similarity between observed images and the model; a motion model which aims to locate the target between consecutive frames over time; and a search strategy which finds the most likely state in the current frame. In our work, we focus attention on designing a robust appearance model to achieve reliable tracking.…”

Section: Introductionmentioning

confidence: 99%

Robust Visual Tracking via Appearance Modeling and Sparse Representation

Li¹,

Fang-lan²,

Nian³

2014

JCP

View full text Add to dashboard Cite

When appearance variation of object, partial occlusion or illumination change in object images occurs, most existing tracking approaches fail to track the target effectively. To deal with the problem, this paper proposed a robust visual tracking method based on appearance modeling and sparse representation. The proposed method exploits two-dimensional principal component analysis (2DPCA) with sparse representation theory for constructing appearance model. Then tracking is achieved by Bayesian inference framework, in which a particle filter is applied to evaluate the target state sequentially over time. In addition, to make the observation model more robust, the incremental learning algorithm is used to update the template set. Both qualitative and quantitative evaluations on four publicly available benchmark video sequences demonstrate that the proposed visual tracking algorithm performs better than several state-of-the-art algorithms.

show abstract

Section: Introductionmentioning

confidence: 99%