Most visual trackers focus on short-term tracking. The target is always in the camera field of view or slight occlusion (OCC). Compared with short-term tracking, long-term tracking is a more challenging task. It requires the ability to capture the target in long-term sequences and undergo frequent disappearances and reappearances of target. Therefore, long-term tracking is much closer to a realistic tracking system. However, few long-term tracking algorithms have been developed and few promising performances have been shown until now. We focus on a long-term visual tracking framework based on parts correlation filters (CFs). Our long-term tracking framework is composed of a part-based short-term tracker and a re-detection module. First, multiple CFs have been applied to locate the target collaboratively and address the partial OCC issue. Second, our method updates the part adaptively based on its motion similarity and reliability score to retain its robustness. Third, a switching strategy has been designed to dynamically activate the re-detection module and interact the search mode between local and global search. In addition, our re-detector is trained by sampling positive and negative samples around the reliable tracking target to adapt to the appearance changes. To evaluate the candidates from the re-detection module, verification has been carried out, which could ensure the precision of recovery. Numerous experimental results demonstrate that our proposed tracking method performs favorably against state-of-the-art methods in terms of accuracy and robustness.
Visual tracking plays a significant role in computer vision. Although numerous tracking algorithms have shown promising results, target tracking remains a challenging task due to appearance changes caused by deformation, scale variation, and partial occlusion. Part-based methods have great potential in addressing the deformation and partial occlusion issues. Owing to the addition of multiple part trackers, most of these part-based trackers cannot run in real time. Correlation filters have been used in target tracking owing to their high efficiency. However, the correlation filter-based trackers face great problems dealing with occlusion, deformation, and scale variation. To better address the above-mentioned issues, we present a scale adaptive part-based tracking method using multiple correlation filters. Our proposed method utilizes the scale-adaptive tracker for both root and parts. The target location is determined by the responses of root tracker and part trackers collaboratively. To estimate the target scale more precisely, the root scale and each part scale are predicted with the sequential Monte Carlo framework. An adaptive weight joint confidence map is acquired by assigning proper weights to independent confidence maps. Experimental results on the publicly available OTB100 dataset demonstrate that our approach outperforms other state-of-the-art trackers.
Target tracking is one of the most topic-active research and also the most important part in the field of computer vision. The typical deformable model target tracking algorithm decomposes each target into multi-sub-blocks, and computes the similarity of both the local areas of each target and the spatial location among each sub-block. However, these algorithms define the area and the number of sub-blocks manually. In the practical application, the tracking system can provide the interaction to select the tracking target real-timely. But it’s difficult to provide the interaction to select the sub-blocks. It means the selection of sub-blocks manually has limitation in the practical application. Aimed at the problems mentioned, this paper presents a method for automatic sub-blocks segmentation. The proposed method integrates the local contrast and the richness of texture details to get a measure function of sub-blocks. Saliency detection based on visual attention model was used to extract salient local contrast. The edge direction dispersion has been used to describe the richness of texture details. Then, the discrimination of each pixel in the target will be computed by the mentioned methods above. Finally, sub-blocks with high discrimination will be chosen for tracking. Experimental results show that the method proposed can achieve more tracking precision compared with the current deformable target tracking algorithm which selected the sub-blocks manually
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.