TbD

所提出的 TbD 跟踪器在一个新创建的视频数据集上进行了评估，该数据集由高速摄像机使用一种新颖的 Trajectory-IoU 度量获得，该数据集推广了传统的 Intersection over Union 并测量帧内轨迹的准确性。所提出的方法在召回和轨迹准确性方面都优于基线。

The proposed TbD tracker was evaluated on a newly created video dataset captured by high-speed cameras with a novel Trajectory-IoU metric. This metric generalizes the traditional Intersection over Union (IoU) and measures the accuracy of intra-frame trajectories. The proposed method outperforms the baseline models in terms of both recall and trajectory accuracy.

搜集汇总

背景与挑战

背景概述

TbD是一个用于视频超分辨率和对象跟踪的数据集，由捷克科学院和捷克理工大学于2019年发布，采用高速摄像机捕获和Trajectory-IoU度量评估帧内轨迹准确性。