作者：Yuang Zhang, Tiancai Wang, Xiangyu Zhang

**作者单位：**上海交通大学、旷视科技、北京人工智能研究院

发布时间：2023

发布期刊/会议：2023 CVPR

出版商：

**论文全称：**MOTRv2: Bootstrapping End-to-End Multi-Object Tracking by Pretrained Object Detectors

论文地址：

CVPR 2023 Open Access Repository

论文代码：

https://github.com/megvii-research/MOTRv2

地位：

个人理解

• **为什么：**现有的端到端方法，如MOTR[43]和TrackFormer[20]，主要由于其检测性能较差而不如其它tracking-by-detection的方法
• **怎么做：**加入一个额外的目标检测器来改进MOTR
  • 将 MOTR 中 detect queries 的可学习位置编码(PE)改为anchor的正余弦位置编码 [30] ，生成了一个基于anchor的 MOTR 跟踪器。利用这种基于anchor的建模方法，通过一个额外的目标检测器生成的 proposal 可以作为 MOTR 的anchor初始化，提供局部先验信息。
  • 利用transformer解码器对 anchor 的相对偏移量（坐标和宽高）进行预测，使检测任务的优化更加容易。

一、摘要

In this paper, we propose MOTRv2, a simple yet effective pipeline to bootstrap end-to-end multi-object tracking with a pretrained object detector. Existing end-to-end methods, e.g. MOTR [43] and TrackFormer [20] are inferior to their tracking-by-detection counterparts mainly due to their poor detection performance. We aim to improve MOTR by elegantly incorporating an extra object detector. We first adopt the anchor formulation of queries and then use an extra object detector to generate proposals as anchors, providing detection prior to MOTR. The simple modification greatly eases the conflict between joint learning detection and association tasks in MOTR. MOTRv2 keeps the query propogation feature and scales well on large-scale benchmarks. MOTRv2 ranks the 1st place (73.4% HOTA on DanceTrack) in the 1st Multiple People Tracking in Group Dance Challenge. Moreover, MOTRv2 reaches state-of-the-art performance on the BDD100K dataset. We hope this simple and effective pipeline can provide some new insights to the endto-end MOT community

• 作者提出了MOTRv2，一个简单而有效的pipline，通过预训练的目标检测器引导端到端的多目标跟踪
  • 采用 anchor formulation of queries，anchor 则通过一个额外的目标检测器来生成
• MOTRv2在第一届 Multiple People Tracking in Group Dance挑战赛中排名第一（73.4% HOTA on DanceTrack）

二、Method