search

maturk / dn-splatter

DN-Splatter + AGS-Mesh: Depth and Normal Priors for Gaussian Splatting
https://maturk.github.io/dn-splatter/
Apache License 2.0
507 stars 32 forks source link

Pool Mesh #57

Open ddenglina opened 4 months ago

ddenglina commented 4 months ago

Hi! Thanks for this awesome project. I trying to train dnsplatter，using custom data collected with an Orbber femto camera. Contain RGB and depth images.

And then I used Spectacular AI SDK processed data.

Finally, I run ns-train dn-splatter --pipeline.model.use-depth-loss True --pipeline.model.sensor-depth-lambda 0.2 --pipeline.model.use-depth-smooth-loss True --pipeline.model.use-sparse-loss True normal-nerfstudio --data Orbber_femto/ --load-normals False ``

gs-mesh dn --load-config outputs/unnamed/dn-splatter/2024-07-31_032623/config.yml --output-dir outputs/unnamed/dn-splatter/2024-07-31_032623/possion_tsdf

The mesh result is so bad. image

Whether I need to set any parameters, I have not changed them. Looking forward to your reply！

maturk commented 4 months ago

Hi @ddenglina, I have not tried dn-splatter with Orbber femto camera before. Can you send me the processed dataset so I can have a try? (google drive or something maybe). Thanks.

ddenglina commented 4 months ago

Hi @ddenglina, I have not tried dn-splatter with Orbber femto camera before. Can you send me the processed dataset so I can have a try? (google drive or something maybe). Thanks.

Thanks you reply. The link is processed dataset by Spectacular AI SDK .Orbber femto data.tar.gz

Looking forward to your reply！