Hi, thanks for releasing the awesome work.
Is it possible to learn the template mesh from multiple frames? Following the same process, we can reconstruct mesh by Volume SDF for each frame and use the root-finding to transfer from live to cano. Does the code include this function? Can we simply set frame_range to multiple frames in the template.yaml?
Only fitting to one frame with A-pose is inaccurate for my case, cause the two legs are too close. And any advice to solve the issue?
Thanks again.
Hi, thanks for releasing the awesome work.
Is it possible to learn the template mesh from multiple frames? Following the same process, we can reconstruct mesh by Volume SDF for each frame and use the root-finding to transfer from live to cano. Does the code include this function? Can we simply set frame_range to multiple frames in the template.yaml?
Only fitting to one frame with A-pose is inaccurate for my case, cause the two legs are too close. And any advice to solve the issue?
Thanks again.