zhang-pengyu / dut-vtuav Goto Github PK

您好作者，我看到您的数据采样是使用大疆禅思系列H20T,我想问问有没有什么设备比较便宜的却又能做到红外图像与可见光图像的同时采集。发现一些海康微视红外的手持设备只能做到记录一种图像。

Thanks for your work! Can you provide us a link of the full dataset for downloading? Thanks again.

Hello, I want to ask you how you generate dense bounding box annotations? Thank you!

Thank you for your contribution to the field of RGB-T tracking! I would like to ask whether the results you evaluated are the results after training on the data set you proposed? If there is training, are sparse boxes or dense boxes used for training?

您好，我是对您这个数据集感兴趣的小白之一，请问标签文件四个坐标的具体含义是什么？

在我的统计中，short-term的测试集总共有63000帧左右可用帧，而short-term训练集只有40000帧左右的可用帧，这正常吗，一般训练集的帧数小于测试集的帧数，这合理吗

师兄你好！我在使用数据集时发现label的长度与序列中图片数对不上，请问VTUAV是每隔10帧标注一次吗？

I notice that your work uses test set to validate some trackers, but you also emphasize this benchmark provides only sparse annotation. So a problem is that did you validate trackers in sparse annotation or use a SOTA tracker to get dense annotation and then do evaluation like TrackingNet?
Thank you.