Search

Correlation filter based tracking algorithms have received significant attention in visual tracking due to their accuracy and speed. However, it is susceptible to distractors and performs poor when it is faced with occlusion and severe deformation. In this paper, we introduce a distractor-aware long-term correlation tracking to distinguish the target and the background. Firstly, the information entropy weighted feature is utilized to deal with different scenarios dynamically. Secondly, in order to facilitate the tracker to distinguish the target from the distractors which lead to inaccurate localization, we propose a novel distractor-aware mask processing. We sample distractor patches and incorporate a binary mask to attenuate the possible negative influences. Thirdly, an improved fast scale searching scheme based on dichotomy is incorporated to provide a tradeoff between the accuracy and efficiency. Finally, the localization ambiguity is detected by using a new criterion based on the variance. It is concluded that introducing the re-detection technique and solving the occlusion play a critical role to facilitate the tracking progress. Furthermore, an adaptive model update strategy is utilized to update the long-term model for alleviating the model drift. Experiments on benchmark database demonstrate that the proposed method achieves good performance compared with other trackers and the strategy can be applied to other trackers.

Correlation filter (CF) based tracking algorithms have tremendously contributed to the field of visual tracking due to the high computational efficiency and competitive performance. Nonetheless, most CF-based trackers are vulnerable to the influence of occlusion and boundary effect, which results in suboptimal performance. In this article, we propose saliency guided visual tracking via correlation filter with log-Gabor filter to robustify its performance under occlusion and boundary effect challenges. Firstly, we propose the CF with log-Gabor filter to get a robust appearance model. The log-Gabor filter is adopted to preprocess the sequence to gain the log-Gabor feature, which provides important cues for tracking since it encodes the texture information. Secondly, considering the prior information, we embed the novel saliency guided adaptive spatial feature selection to filter learning to preserve the spatial structure in the lower manifold and mitigate boundary distortion. Thirdly, the occlusion estimating strategy, performing on-line evaluation of tracking, triggers the motion estimation module to optimize the optimal location. Experiments on benchmark databases demonstrate the enhanced discrimination and interpretability of the proposed tracker and its superiority over other trackers.