15 papers
Epipolar Geometry Improves Video Generation Models
Orest Kupyn, Théo Uscidda, Marta Tintore Gazulla +3
Video generation models have advanced significantly through the latent diffusion transformers trained with rectified flow techniques. Yet these models still struggle with geometric…
SA4Depth: Consistent Pose-Depth Scale Alignment for Self-Supervised Monocular Depth Estimation
Changxuan Li, Nadine Berner, Nassir Navab +2
Self-supervised depth estimation from monocular sequences relies on the joint learning of a depth and a pose network. Despite abundant research done to improve the depth network, e…
SING3R-SLAM: Submap-based Indoor Monocular Gaussian SLAM with 3D Reconstruction Priors
Kunyi Li, Michael Niemeyer, Sen Wang +3
Recent advances in dense 3D reconstruction have demonstrated strong capability in accurately capturing local geometry. However, extending these methods to incremental global recons…
SuperGSeg: Open-Vocabulary 3D Segmentation with Structured Super-Gaussians
Siyun Liang, Sen Wang, Kunyi Li +5
3D Gaussian Splatting has recently gained traction for its efficient training and real-time rendering. While its vanilla representation is mainly designed for view synthesis, recen…
Language-Guided Open-World Anomaly Segmentation
Klara Reichard, Nikolas Brasch, Nassir Navab +1
Open-world and anomaly segmentation methods seek to enable autonomous driving systems to detect and segment both known and unknown objects in real-world scenes. However, existing m…
MonoGSDF: Exploring Monocular Geometric Cues for Gaussian Splatting-Guided Implicit Surface Reconstruction
Kunyi Li, Michael Niemeyer, Zeyu Chen +2
Accurate meshing from monocular images remains a key challenge in 3D vision. While state-of-the-art 3D Gaussian Splatting (3DGS) methods excel at synthesizing photorealistic novel…