VOT2018

The VOT2018 dataset contains 60 videos and 100 frames for visual object tracking.

BibTex: