Spatially Attentive Correlation Filters for Visual Tracking

Huai Qin, Zhixiong Pi, Changqian Yu, Changxin Gao, Jin-Gang Yu, Nong Sang

Abstract

Although correlation filter based trackers have recently demonstrated excellent performance, they still suffer from the boundary effects. The cosine window is introduced to alleviate the boundary affects, which however may result in poor performance in case of occlusion or fast motion. To address this problem, we propose a simple yet effective framework, which builds a spatially attentive model with multiple features to guide the detection of the correlation filter based trackers. The proposed method not only can breakthrough the spatial extent of cosine window, but also can provides prior information about the target object. Moreover, to model a robust object prior, we propose a generic strategy for adaptive fusion and update of multiple features. Extensive experiments over multiple tracking benchmarks demonstrate the superior accuracy and real-time performance of our methods compared to the state-of-the-art trackers.
BibTeX
@inproceedings{qin2018spatially,
  author = {Huai Qin and Zhixiong Pi and Changqian Yu and Changxin Gao and Jin-Gang Yu and Nong Sang},
  title = {{Spatially Attentive Correlation Filters for Visual Tracking}},
  booktitle = {ICIP},
  year = {2018},
  doi = {10.1109/icip.2018.8451045},
}

Loading related content...

Figure