Open sfchen94 opened 1 year ago
Hi, are you using our tracking/analysis_results.py to test? I just retested raw results and the performance are the same as in the paper.
@chenxin-dlut
Hi, I guess the issue might be related to the format conversion of the ground truth.
Based on my conversion approach, I have been able to achieve the same performance as reported in other papers. However, SeqTrack is an exception to this. Therefore, I assume you converted it in your own way.
Could you please provide the '_pytracking.txt' data that you converted?
Hi, have you solved it? I alse need the 30fps annotation.
Hi, I have evaluated your raw results on the NFS dataset, and it appears to perform worse than the reported results in the paper, with a difference of approximately -0.8%. Could you please confirm this observation?