Visualizing multiple people in the same world frame

Tested methods: WHAM, GVHMR, TRACE

TL;DR

그나마 GVHMR이 괜찮음.

They all give poor results. They don't have ground estimation. I made issues in each repo to clarify whether I am doing something wrong or it is the fundamental limitation of their methods. (Oct 12th, 2024)

https://github.com/yohanshin/WHAM/issues/118

Fail to put multiple persons in the same world frame · Issue #118 · yohanshin/WHAM

https://github.com/zju3dv/GVHMR/issues/29

Fail to put multiple persons in the same world frame · Issue #29 · zju3dv/GVHMR

https://github.com/Arthur151/ROMP/issues/517

Failure case of TRACE in PoseTrack video · Issue #517 · Arthur151/ROMP

TRACE 이거는 논문 contribution과 완전 배치되는 결과이고, 코드 까보면 논문 주장과 다른게 있어서, 내 추측과 visualization이 맞다면 좀 심각한 문제 같다. 그럴 일이야 없겠지만 논문 철회해야 하는 수준인 거 같은데. 내가 틀린 거이길.