search

NVlabs / GLAMR

[CVPR 2022 Oral] Official PyTorch Implementation of "GLAMR: Global Occlusion-Aware Human Mesh Recovery with Dynamic Cameras”.
https://nvlabs.github.io/GLAMR
Other
359 stars 33 forks source link

Support for multiple person #11

Closed haofanwang closed 2 years ago

haofanwang commented 2 years ago

Is there any plan?

Khrylx commented 2 years ago

This mainly involves integrating robust multi-object tracking into the demo. Due to limited bandwidth, there is no current plan for this. Any community efforts would be much appreciated.

haofanwang commented 2 years ago

I have added a PR https://github.com/NVlabs/GLAMR/pull/12 to support it.

Khrylx commented 2 years ago

Thank you so much, @haofanwang!

Khrylx commented 2 years ago

quick question: is this part redundant? https://github.com/NVlabs/GLAMR/blob/463dda1c23cdce70c1e0a19a835fdbfa45777be3/global_recon/run_demo.py#L80-L83

haofanwang commented 2 years ago

I see, these lines are not necessary (just orders the keys in est_dict) and should be removed.

Khrylx commented 2 years ago

will there be a problem if the orders are not maintained? If so, we can probably add it to the demo.py

haofanwang commented 2 years ago

It shouldn't be a problem, I have tested and the visualization results are same.

Khrylx commented 2 years ago

ok, thanks!