CVPR2022_STNet icon indicating copy to clipboard operation
CVPR2022_STNet copied to clipboard

Published 20 hours ago • verified publisher icon Jee-King

Why is the accuracy of the test for repetition lower than that in the paper?

Open HYi0 opened this issue 2 years ago • 1 comments

I replicated the test code on FE240 dataset. In the paper, the success plot is 0.585, but in my tests, the accuracy was only 0.507. I used the pretrained which you provided. So why is this happening?

HYi0 avatar Feb 14 '23 11:02 HYi0

Hi, We evaluate all trackers against pytracking.
The Analysis section in the Readme provides how to evaluate the tracker.

Jee-King avatar Feb 19 '23 02:02 Jee-King