There is a lot of noise around the object reconstructed by neus-acc.

[wujinhu1999]

Hello authors, thank you for your great work! I encountered some problems while training neus-acc. There is a lot of noise around the reconstructed object. How to remove this? Are there any hyperparameters that need to be adjusted? Looking forward to your reply. My training command is as follows:ns-train neus-acc --pipeline.model.sdf-field.inside-outside False --vis wandb --experiment-name neus-acc-dtu24 sdfstudio-data --data data/dtu/scan24

[niujinshuchong]

Hi, these regions correspond to the white table and black background in the input images. Therefore, no correspondence can be found and the optimised surface is kind of random. You can remove it with object masks.

[wujinhu1999]

Hi, these regions correspond to the white table and black background in the input images. Therefore, no correspondence can be found and the optimised surface is kind of random. You can remove it with object masks.

Hello, it means use the command '--include_foreground_mask True'?

[hanjoonwon]

Hi, these regions correspond to the white table and black background in the input images. Therefore, no correspondence can be found and the optimised surface is kind of random. You can remove it with object masks.

Hello, it means use the command '--include_foreground_mask True'?

Can you tell me the train option?
i trained with ns-train neus-facto --pipeline.datamanager.train-num-rays-per-batch 2048 --pipeline.model.sdf-field.use-grid-feature False --pipeline.model.sdf-field.inside-outside False --pipeline.model.background-model mlp --pipeline.model.mono-depth-loss-mult 0.0 --pipeline.model.mono-normal-loss-mult 0.01 --vis wandb --trainer.steps_per_save 5000 --trainer.steps-per-eval-image 5000
but I'm getting very strange results.

[wujinhu1999]

For the neus-facto model, you can just use the author's default commands. My result comes from the neus-acc model.

[hanjoonwon]

For the neus-facto model, you can just use the author's default commands. My result comes from the neus-acc model.

Thank you i don't know why my outputs are so poor