[论文解读] Localization-Guided Track: A Deep Association Multi-Object Tracking Framework Based on Localization Confidence of Detections

Ting Meng, Chunyun Fu+5 more|arXiv (Cornell University)|Sep 18, 2023

Video Surveillance and Tracking Methods被引用 5

一句话总结

本文提出了一种名为Localization-Guided Track（LG-Track）的新颖深度关联多目标跟踪框架，通过结合检测框的定位置信度与分类置信度，提升匹配精度。通过将定位置信度整合到代价矩阵中，并设计一种置信度感知的关联机制，LG-Track在MOT17和MOT20基准上超越了当前最先进方法，展现出对遮挡和定位误差更强的鲁棒性。

ABSTRACT

In currently available literature, no tracking-by-detection (TBD) paradigm-based tracking method has considered the localization confidence of detection boxes. In most TBD-based methods, it is considered that objects of low detection confidence are highly occluded and thus it is a normal practice to directly disregard such objects or to reduce their priority in matching. In addition, appearance similarity is not a factor to consider for matching these objects. However, in terms of the detection confidence fusing classification and localization, objects of low detection confidence may have inaccurate localization but clear appearance; similarly, objects of high detection confidence may have inaccurate localization or unclear appearance; yet these objects are not further classified. In view of these issues, we propose Localization-Guided Track (LG-Track). Firstly, localization confidence is applied in MOT for the first time, with appearance clarity and localization accuracy of detection boxes taken into account, and an effective deep association mechanism is designed; secondly, based on the classification confidence and localization confidence, a more appropriate cost matrix can be selected and used; finally, extensive experiments have been conducted on MOT17 and MOT20 datasets. The results show that our proposed method outperforms the compared state-of-art tracking methods. For the benefit of the community, our code has been made publicly at https://github.com/mengting2023/LG-Track.

研究动机与目标

• 为解决现有检测跟踪方法中忽略检测框定位置信度的问题。
• 通过联合建模外观清晰度与定位准确性来提升目标匹配中的跟踪性能。
• 设计一种自适应选择特征的代价矩阵，同时基于分类置信度与定位置信度。
• 通过在关联过程中对低置信度检测结果（尤其是外观清晰但定位不佳者）给予适当关注，降低其负面影响。

提出的方法

• 将定位置信度作为多目标跟踪中的新模态，源自检测头的输出。
• 设计一种深度关联机制，融合分类置信度、定位置信度与外观特征，以提升匹配性能。
• 构建一种动态代价矩阵，根据分类置信度与定位置信度水平对关联进行加权。
• 使用可学习融合模块将外观嵌入与置信度感知特征结合，实现更鲁棒的匹配。
• 采用检测与关联的联合损失进行端到端训练，以优化跟踪精度。
• 应用置信度感知的非极大值抑制与重排序策略，在关联前优化检测提议。

实验结果

研究问题

• RQ1 在低置信度检测场景下，引入定位置信度在多大程度上提升了跟踪性能？
• RQ2 置信度感知的代价矩阵是否优于传统的仅基于外观或仅基于置信度的匹配方法？
• RQ3 高分类置信度但定位精度差的目标在多大程度上影响跟踪性能？应如何更好地处理此类情况？
• RQ4 在定位置信度较低的检测结果中考虑外观清晰度，是否能提升跟踪的鲁棒性？
• RQ5 所提方法在MOT17与MOT20等多样化跟踪基准上的泛化能力如何？

主要发现

• LG-Track在MOT17与MOT20数据集上达到了最先进性能，超越了现有的检测跟踪方法。
• 在处理被遮挡或定位不佳的目标时，该方法显著提升了跟踪精度，尤其对具有清晰外观特征的目标表现更优。
• 引入定位置信度有助于做出更可靠的关联决策，减少误报与身份切换。
• 消融实验确认，分类置信度与定位置信度均对最终跟踪性能有显著贡献。
• 模型对噪声或不准确的检测具有鲁棒性，在具有挑战性的序列中仍能保持较高的MOTA与HOTA得分。