7 papers
Zero-Shot Depth from Defocus
Yiming Zuo, Hongyu Wen, Venkat Subramanian +5
Depth from Defocus (DfD) is the task of estimating a dense metric depth map from a focus stack. Unlike previous works overfitting to a certain dataset, this paper focuses on the ch…
Telescope: Learnable Hyperbolic Foveation for Ultra-Long-Range Object Detection
Parker Ewen, Dmitriy Rivkin, Mario Bijelic +1
Autonomous highway driving, especially for long-haul heavy trucks, requires detecting objects at long ranges beyond 500 meters to satisfy braking distance requirements at high spee…
ChopGrad: Pixel-Wise Losses for Latent Video Diffusion via Truncated Backpropagation
Dmitriy Rivkin, Parker Ewen, Lili Gao +5
Recent video diffusion models achieve high-quality generation through recurrent frame processing where each frame generation depends on previous frames. However, this recurrent mec…
UniLiPs: Unified LiDAR Pseudo-Labeling with Geometry-Grounded Dynamic Scene Decomposition
Filippo Ghilotti, Samuel Brucker, Nahku Saidy +3
Unlabeled LiDAR logs, in autonomous driving applications, are inherently a gold mine of dense 3D geometry hiding in plain sight - yet they are almost useless without human labels,…
LSD-3D: Large-Scale 3D Driving Scene Generation with Geometry Grounding
Julian Ost, Andrea Ramazzina, Amogh Joshi +3
Large-scale scene data is essential for training and testing in robot learning. Neural reconstruction methods have promised the capability of reconstructing large physically-ground…
SAMFusion: Sensor-Adaptive Multimodal Fusion for 3D Object Detection in Adverse Weather
Edoardo Palladin, Roland Dietze, Praveen Narayanan +2
Multimodal sensor fusion is an essential capability for autonomous robots, enabling object detection and decision-making in the presence of failing or uncertain inputs. While recen…