MOSE: Monocular Semantic Reconstruction Using NeRF-Lifted Noisy Priors
This work addresses the challenge of 3D scene understanding from monocular images, which is important for applications like robotics and augmented reality, but it appears incremental as it builds on existing neural field methods with specific enhancements.
The paper tackles the problem of reconstructing dense, semantically annotated 3D meshes from monocular images by proposing MOSE, a neural field approach that lifts noisy 2D priors to 3D, resulting in improved performance on 3D semantic segmentation, 2D semantic segmentation, and 3D surface reconstruction tasks on the ScanNet dataset.
Accurately reconstructing dense and semantically annotated 3D meshes from monocular images remains a challenging task due to the lack of geometry guidance and imperfect view-dependent 2D priors. Though we have witnessed recent advancements in implicit neural scene representations enabling precise 2D rendering simply from multi-view images, there have been few works addressing 3D scene understanding with monocular priors alone. In this paper, we propose MOSE, a neural field semantic reconstruction approach to lift inferred image-level noisy priors to 3D, producing accurate semantics and geometry in both 3D and 2D space. The key motivation for our method is to leverage generic class-agnostic segment masks as guidance to promote local consistency of rendered semantics during training. With the help of semantics, we further apply a smoothness regularization to texture-less regions for better geometric quality, thus achieving mutual benefits of geometry and semantics. Experiments on the ScanNet dataset show that our MOSE outperforms relevant baselines across all metrics on tasks of 3D semantic segmentation, 2D semantic segmentation and 3D surface reconstruction.