Closed sjtuytc closed 4 years ago
What causes this problem? I doubt that the gt label 'spl' does not correctly indicate the snitch in this case.
Sorry, I'm a bit confused. Why do you say it's not over the snitch? It looks to me that the box is over the snitch in these images.
Just for reference, this is the snitch object: 3 golden toruses intertwined (different from a yellow ball)
Sorry, I make it wrong because in Figure 2 of your paper, the snitch is contained by the tone so I took the tone as the snitch.
Hi, How do i get the same 33.9? I got the following result when running python main.py
Tracking baseline accuracies: {'top_1': 0.03755299818291944, 'top_5': 0.15626892792247124, 'l1_dist': 3.509993943064809}
Tracking baseline accuracy (GT end loc, this should be 1.0): {'top_1': 0.9993943064809206, 'top_5': 0.9993943064809206, 'l1_dist': 0.0036341611144760752}
Also if I understood it correctly, the blue box is the gt for the end position of the snitch and the yellow box is the tracking performance of the DaSiamRPN model?
I did get the results of 33.9.
Hi, when I'm visualizing the results of the tracking method (e.g. turn on DEBUG_STORE_VIDEO), I found that the gt location of spl does not place the snitch, but the ball. Btw, I can reproduce the number 33.9 exactly. You can see in the following figure and videos, the gt (in blue rectangle) refers to the yellow ball, but not the snitch. I check the gt number is the same as shown in the visualization.
cater_debug_tracking_videos.zip