The clothing in the mesh extracted by extract_mesh is incorrect.

[Liu-xiaoyu11]

Hello, @yfeng95 @LeeBeen1：
I'm sorry to bother you again. When I ran python main_demo.py --vis_type extract_mesh --frame_id 0 using my own video, I got an incorrect result as shown below:
,
it's clearly visible that the clothing extracted by NeRF is incorrect. How can this problem be solved? Do you have any good suggestions? I really hope to get your guidance and look forward to your reply.

[LeeBeen1]

Sorry for the late response, did u solve this?

[Liu-xiaoyu11]

Hello, @LeeBeen1 :
Thank you very much for replying to me, now the problem is not solved, the processed video is a video shot by myself (cell phone shooting, 1080*1920, lighting is more random, the background is not completely clean), retrained python main_train.py --data_cfg configs/data/mpiis/DSC_7157. yml --exp_cfg configs/exp/stage_1_hybrid_perceptual.yml, got this hybrid corresponding model.tml. yml --exp_cfg configs/exp/stage_1_hybrid_perceptual.yml to get the corresponding model.tar for this hybrid, but the clothing in the mesh extracted by extract_mesh is incorrect, as shown below:

what's wrong with this, do you have any suggestions, looking forward to getting your reply earlier, thanks a lot.
Also, the video I shot in the field was not processed well, as shown below：