Alpha-Refine: Boosting Tracking Performance by Precise Bounding Box Estimation
This work addresses the need for more accurate and transferable refinement in visual tracking, which is incremental as it builds upon existing multi-stage strategies.
The authors tackled the problem of limited transferability and precision in existing refinement modules for visual tracking by proposing Alpha-Refine, a novel module that improves tracking performance on benchmarks like TrackingNet, LaSOT, and VOT2018 when applied to five state-of-the-art base trackers.
In recent years, the multiple-stage strategy has become a popular trend for visual tracking. This strategy first utilizes a base tracker to coarsely locate the target and then exploits a refinement module to obtain more accurate results. However, existing refinement modules suffer from the limited transferability and precision. In this work, we propose a novel, flexible and accurate refinement module called Alpha-Refine, which exploits a precise pixel-wise correlation layer together with a spatial-aware non-local layer to fuse features and can predict three complementary outputs: bounding box, corners and mask. To wisely choose the most adequate output, we also design a light-weight branch selector module. We apply the proposed Alpha-Refine module to five famous and state-of-the-art base trackers: DiMP, ATOM, SiamRPN++, RTMDNet and ECO. The comprehensive experiments on TrackingNet, LaSOT and VOT2018 benchmarks demonstrate that our approach significantly improves the tracking performance in comparison with other existing refinement methods. The source codes will be available at https://github.com/MasterBin-IIAU/AlphaRefine.