Get 3D scene by preferring walls and floor to be flat, vertical, or horizontal with FAWN
Get 3D scene by preferring walls and floor to be flat, vertical, or horizontal with FAWN
FAWN: Floor-And-Walls Normal Regularization for Direct Neural TSDF Reconstruction
arXiv paper abstract https://arxiv.org/abs/2406.12054
arXiv PDF paper https://arxiv.org/pdf/2406.12054
Leveraging 3D semantics for direct 3D reconstruction has a great potential yet unleashed.
For instance, by assuming that walls are vertical, and a floor is planar and horizontal, … can correct distorted room shapes and eliminate local artifacts such as holes, pits, and hills.
… propose FAWN, a modification of truncated signed distance function (TSDF) reconstruction methods, which … detecting walls and floor .. and penalizing … normals for deviating from the horizontal and vertical …
Implemented as a 3D sparse convolutional module, FAWN can be incorporated into any trainable pipeline that predicts TSDF.
Since FAWN requires 3D semantics only for training, no additional limitations … are imposed … demonstrate, that FAWN-modified methods use semantics more effectively
… apply … modification to state-of-the-art TSDF reconstruction methods, and demonstrate a quality gain in SCANNET, ICL-NUIM, TUM RGB-D, and 7SCENES benchmarks.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
