论文标题
Giaotracker：MCMOT的综合框架，具有全球信息并优化Visdrone 2021
GIAOTracker: A comprehensive framework for MCMOT with global information and optimizing strategies in VisDrone 2021
论文作者
论文摘要
近年来，多个对象跟踪任务的算法从深度模型和视频质量中的进步中受益。但是，在诸如无人机视频之类的具有挑战性的场景中，它们仍然遇到问题，例如小物体，摄像头动作和查看更改。在本文中，我们提出了一个新的多重对象跟踪器，该跟踪器采用全局信息和一些优化策略，名为Giaotracker。它包括三个阶段，即在线跟踪，全球链接和后处理。给定的每个帧检测，第一阶段使用相机运动，对象运动和对象外观的信息生成可靠的轨迹。然后，它们通过利用全球线索并通过四种后处理方法来完善轨迹。凭借这三个阶段的有效性，Giaotracker在Visdrone MOT数据集中取得了最新的性能，并在Visdrone2021 MOT挑战中赢得了第三名。
In recent years, algorithms for multiple object tracking tasks have benefited from great progresses in deep models and video quality. However, in challenging scenarios like drone videos, they still suffer from problems, such as small objects, camera movements and view changes. In this paper, we propose a new multiple object tracker, which employs Global Information And some Optimizing strategies, named GIAOTracker. It consists of three stages, i.e., online tracking, global link and post-processing. Given detections in every frame, the first stage generates reliable tracklets using information of camera motion, object motion and object appearance. Then they are associated into trajectories by exploiting global clues and refined through four post-processing methods. With the effectiveness of the three stages, GIAOTracker achieves state-of-the-art performance on the VisDrone MOT dataset and wins the 3rd place in the VisDrone2021 MOT Challenge.
