DN-Splatter: Depth and Normal Priors for Gaussian Splatting and Meshing

Matias Turkulainen, Xuqian Ren, Iaroslav Melekhov, Otto Seiskari, Esa Rahtu, Juho Kannala

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference article in proceedingsScientificvertaisarvioitu

Abstrakti

High-fidelity 3D reconstruction of common indoor scenes is crucial for VR and AR applications. 3D Gaussian splatting, a novel differentiable rendering technique, has achieved state-of-the-art novel view synthesis results with high rendering speeds and relatively low training times. However, its performance on scenes commonly seen in indoor datasets is poor due to the lack of geometric constraints during optimization. In this work, we explore the use of readily accessible geometric cues to enhance Gaussian splatting optimization in challenging, ill-posed, and textureless scenes. We extend 3D Gaussian splatting with depth and normal cues to tackle challenging indoor datasets and showcase techniques for efficient mesh extraction. Specifically, we regularize the optimization procedure with depth information, enforce local smoothness of nearby Gaussians, and use off-the-shelf monocular networks to achieve better alignment with the true scene geometry. We propose an adaptive depth loss based on the gradient of color images, improving depth estimation and novel view synthesis results over various baselines. Our simple yet effective regularization technique enables direct mesh extraction from the Gaussian representation, yielding more physically accurate reconstructions of indoor scenes.
AlkuperäiskieliEnglanti
OtsikkoWinter Conference on Applications of Computer Vision
AlaotsikkoWACV 2025
TilaHyväksytty/In press - 2025
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisussa
TapahtumaIEEE Winter Conference on Applications of Computer Vision - Tucson, Yhdysvallat
Kesto: 28 helmik. 20254 maalisk. 2025

Conference

ConferenceIEEE Winter Conference on Applications of Computer Vision
LyhennettäWACV
Maa/AlueYhdysvallat
KaupunkiTucson
Ajanjakso28/02/202504/03/2025

Sormenjälki

Sukella tutkimusaiheisiin 'DN-Splatter: Depth and Normal Priors for Gaussian Splatting and Meshing'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä