Indoor 3D scene reconstruction 10x faster learn and 100x faster render using sparse voxels with Dong
--
Indoor 3D scene reconstruction 10x faster learn and 100x faster render using sparse voxels with Dong
Fast Monocular Scene Reconstruction with Global-Sparse Local-Dense Grids
arXiv paper abstract https://arxiv.org/abs/2305.13220
arXiv PDF paper https://arxiv.org/pdf/2305.13220.pdf
Indoor scene reconstruction from monocular images … advances in neural field representations and monocular priors have led to remarkable results in scene-level surface reconstructions.
… propose to directly use signed distance function (SDF) in sparse voxel block grids for fast and accurate scene reconstruction
… globally sparse and locally dense data structure exploits surfaces’ spatial sparsity, enables cache-friendly queries, and allows direct extensions to multi-modal data such as color and semantic labels.
… develop a scale calibration algorithm for fast geometric initialization from monocular depth priors … apply differentiable volume rendering from this initialization to refine details with fast convergence.
… introduce efficient high-dimensional Continuous Random Fields (CRFs) to further exploit the semantic-geometry consistency between scene objects.
… approach is 10x faster in training and 100x faster in rendering while achieving comparable accuracy to state-of-the-art neural implicit methods.
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
