MVTrack
收藏资源简介:
MVTrack数据集是由上海科技大学生物医学工程学院和国家重点实验室构建的多视角跟踪数据集,包含234K个高质量标注帧,涵盖27个不同对象类别,跨越9个具有挑战性的跟踪属性,如遮挡和形变。该数据集由3-4个校准相机捕获,提供精确的边界框(BBox)注释,并包含训练、验证和测试集,是首个为训练类无关多视角跟踪方法提供的全面基准数据集,也为评估这些方法提供了丰富的手段。
The MVTrack dataset is a multi-view tracking benchmark constructed by the School of Biomedical Engineering, ShanghaiTech University and the State Key Laboratory. It contains 234K high-quality annotated frames, covering 27 distinct object categories and spanning 9 challenging tracking attributes such as occlusion and deformation. Captured by 3-4 calibrated cameras, it provides precise bounding box (BBox) annotations, and includes training, validation, and test splits. It is the first comprehensive benchmark dataset for training class-agnostic multi-view tracking methods, and also provides a rich means for evaluating such methods.

- 1MITracker: Multi-View Integration for Visual Object Tracking上海科技大学 生物医学工程学院 & 国家重点实验室 · 2025年



