Closed kumatheworld closed 3 years ago
Sorry, I've just found the following lines in the paper.
Coarse annotations are provided by YT-BB at 1 fps. In order to increase the annotation density, we rely on a mixture of state-of-the-art trackers to fill in missing annotations. We claim that any tracker is reliable on a small time lapse of 1 second. We present in Section 4 the performance of state-of-the-art trackers on 1 second-long video segments from OTB100. As a result, we densely annotated the 30,132 videos using a weighted average between a forward and a backward pass using the DCF tracker [16]. By doing so, we provide a densely annotated training dataset for object tracking, along with code for automatically downloading videos from YouTube and extracting the annotated frames.
It seems like that the negative coordinates come from the tracker outputs.
I found many ground truth boxes sticking out of the FOV.
For example, If you look at the first few lines of TRAIN_0/anno/-3TIfnTSM6c_2.txt, you can find negative values in the first column as below.
Is this intended?