LGTrack: Exploiting Local and Global Properties for Robust Visual Tracking

Re-detection is a necessary capability for long-term tracking. Target candidate proposals in the whole image can provide a chance of tracking reset when tracking fails due to tracking drift or target invisibility. In this paper, we propose a unified local-global tracker based on the same transformer architecture sharing weights, which can not only search in a continuous local region but also provide target candidates of the global image in every frame. The requirements of both long-term and short-term scenarios can be addressed using a unified model. A simple proposal selection scheme is adopted to properly select the candidate proposals of re-detection, to assist tracking and obtain better performance. The scheme performs re-evaluation of all high-quality proposals based on a transformer-based embedding network, once the predicted state of the local tracking is not sufficient to be accurate. To capture appearance variations brought by online updates in minimum risks, a long-term-friendly dynamic template update scheme is also designed. Extensive experiments are conducted to demonstrate the effectiveness of our proposed tracker, including three short-term tracking benchmarks and six long-term benchmarks. Our tracker can achieve results comparable to that of the state-of-the-art. The proposed tracker can also work well in balancing the performance and speed, achieving an average speed of approximately 25 fps tested on LaSOT testing set.

I. Introduction

Visual object tracking aims to localize the specified object in a sequence given its initial state in the first frame. With the continuous innovation of the feature extraction backbone [1], [2], [3] and the manner of relation modeling [4], [5], [6], [7], [8], [9] between template and search regions, and the exploration on the saving of annotation costs [10], [11], the performance of tracking algorithms has achieved new heights. Meanwhile, the long-term tracking has gradually attracted more attention. The difference from short-term tracking, which has always been paid the most attention, is that in long-term tracking, the video duration is longer, and target disappearances and reappearances may occur more frequently. The long sequences can bring more complicated challenges and more obvious tracking error accumulation. Target disappearances and reappearances lead to the necessity of the re-detection capability of trackers. For the short-term tracking, sequences with targets always visible and a short duration are dominant. Therefore, searching in a local region based on the target location in the previous frame can handle most scenarios.