About the results of 'mvpose' on Panoptic dataset

[Billccx]

Hi ! Your work is truly impressive!
I noticed that the quantified evaluation results of the MVPose algorithm on the Panoptic dataset are very poor. In the MVPose paper, it is mentioned that only qualitative analysis was conducted on the Panoptic dataset. Could you please tell me how you obtained this result? Was it through experiments or referenced from other literature? If you obtained this result through experiments, could you consider open-sourcing the relevant experimental code?

[DanBmh]

Hi, I got quite similar results in my own experiments.
The frame sampling approach was different, so the results are slightly different as well.
You can find my results (including code) here:
https://gitlab.com/Percipiote/MVPoseWrapper/-/blob/master/RESULTS.md?ref_type=heads#panoptic