Open chuxiang93 opened 1 year ago
Describe the bug Bad reconstruction result on my custom dataset. The reconstructed target is less than half the size.
2d image: mesh result:
To Reproduce Steps to reproduce the behavior: 1,dataset conversion python scripts/datasets/process_nerfstudio_to_sdfstudio.py --data-type "colmap" --scene-type "object" --data data/nerfstudio/toy --output-dir data/sdfstudio/toy --mono-prior --omnidata-path ../workspace/omnidata/omnidata_tools/torch/ --pretrained-models ../workspace/omnidata/omnidata_tools/torch/pretrained_models/
python scripts/datasets/process_nerfstudio_to_sdfstudio.py --data-type "colmap" --scene-type "object" --data data/nerfstudio/toy --output-dir data/sdfstudio/toy --mono-prior --omnidata-path ../workspace/omnidata/omnidata_tools/torch/ --pretrained-models ../workspace/omnidata/omnidata_tools/torch/pretrained_models/
2,train code ns-train neus-facto --pipeline.model.sdf-field.inside-outside False --pipeline.model.sdf-field.bias 0.3 --pipeline.model.background-model mlp --pipeline.model.mono-depth-loss-mult 0.1 --pipeline.model.mono-normal-loss-mult 0.05 --experiment-name neus-facto-toy sdfstudio-data --data data/sdfstudio/toy --include_mono_prior True
ns-train neus-facto --pipeline.model.sdf-field.inside-outside False --pipeline.model.sdf-field.bias 0.3 --pipeline.model.background-model mlp --pipeline.model.mono-depth-loss-mult 0.1 --pipeline.model.mono-normal-loss-mult 0.05 --experiment-name neus-facto-toy sdfstudio-data --data data/sdfstudio/toy --include_mono_prior True
Have you resolved this issue?
I also encountered this problem, how did you solve it?
Describe the bug Bad reconstruction result on my custom dataset. The reconstructed target is less than half the size.
2d image: mesh result:
To Reproduce Steps to reproduce the behavior: 1,dataset conversion
python scripts/datasets/process_nerfstudio_to_sdfstudio.py --data-type "colmap" --scene-type "object" --data data/nerfstudio/toy --output-dir data/sdfstudio/toy --mono-prior --omnidata-path ../workspace/omnidata/omnidata_tools/torch/ --pretrained-models ../workspace/omnidata/omnidata_tools/torch/pretrained_models/
2,train code
ns-train neus-facto --pipeline.model.sdf-field.inside-outside False --pipeline.model.sdf-field.bias 0.3 --pipeline.model.background-model mlp --pipeline.model.mono-depth-loss-mult 0.1 --pipeline.model.mono-normal-loss-mult 0.05 --experiment-name neus-facto-toy sdfstudio-data --data data/sdfstudio/toy --include_mono_prior True