SOURCE-LINKED INTELLIGENCE

Geometry Without Coordinates: LiDAR Diffusion as a 3D Feature Bridge

arXiv · AI, language, vision and robotics · article · Sep 9, 2026 · UTC

Transferring the rich priors of large 2D foundation models to sparse 3D LiDAR remains challenging, as training native 3D foundation models at comparable scale is limited by data and annotation scarcity. We introduce a LiDAR-conditioned diffusion model trained on pseudo-labels from off-the-shelf 2D foundation models. The model supports multiple output modalities, including depth, semantic segmentation and instance prediction, selectable via a textual task prompt. Because the model is conditioned on LiDAR, both its outputs and its intermediate UNet features can be projected back onto the input p

Read original source ↗ Open in workspace

recordType

paper

region

Global