I download renderpeople dataset and run bash eval_renderpeople_512x512.sh. What puzzles me is that renderpeople dataset doesn't contain SMPL parameters, so according to SHERF framework, how can it apply LBS to transform points from the canonical space to target/observation spaces, and synthesis noval pose?
Maybe I have some misconceptions, could you help me?
I download renderpeople dataset and run
bash eval_renderpeople_512x512.sh
. What puzzles me is that renderpeople dataset doesn't contain SMPL parameters, so according to SHERF framework, how can it apply LBS to transform points from the canonical space to target/observation spaces, and synthesis noval pose? Maybe I have some misconceptions, could you help me?Thanks a lot.