Until now, enabling depth loss multi always gave me bad results. I don't know what else to do, but at least neus gives good quality.
However, when I change only the method to something else in the above options and run it, the result is too bad, why?
++++I gave the object a clear acrylic base underneath in order to 3D recon the underside of the object.
The base is also reconstructed, so when the base is removed, the area under the battery is empty.
Would it be more helpful to not have the transparent base for a better reconstruction of the bottom?
https://drive.google.com/drive/folders/1Q9zSM8sCsQ4n-6QKl4GEujuznx-abxvh?usp=drive_link
i trained with
ns-train neus --pipeline.datamanager.train-num-rays-per-batch 2048 --pipeline.model.sdf-field.inside-outside False\ --pipeline.model.sdf-field.hidden-dim 256 --pipeline.model.sdf-field.num-layers 2 \ --pipeline.model.sdf-field.num-layers-color 2 --pipeline.model.sdf-field.use-grid-feature False\ --pipeline.model.sdf-field.bias 0.3 --pipeline.model.sdf-field.beta-init 0.3 --pipeline.model.background-model mlp \ --pipeline.model.sdf-field.use-appearance-embedding True\ --trainer.steps-per-eval-image 5000 --trainer.max-num-iterations 50000 --viewer.websocket-port 7008 \ --pipeline.model.near-plane 0.05 --pipeline.model.far-plane 2. --pipeline.model.overwrite-near-far-plane True\ --pipeline.model.mono-normal-loss-mult 0.01 --pipeline.model.mono-depth-loss-mult 0.0\
Until now, enabling depth loss multi always gave me bad results. I don't know what else to do, but at least neus gives good quality.
However, when I change only the method to something else in the above options and run it, the result is too bad, why?
++++I gave the object a clear acrylic base underneath in order to 3D recon the underside of the object. The base is also reconstructed, so when the base is removed, the area under the battery is empty. Would it be more helpful to not have the transparent base for a better reconstruction of the bottom?