LiFi: LiDAR Generation from Multi-View Images via Geometric and Semantic Collaborative Guidance

Key points

• In this work, we introduce LiFi (LiDAR From images), a dedicated framework for generating visually and geometrically aligned LiDAR scenes from multi-view camera images.1
• The geometry branch constructs scene features from images through a reverse sampling algorithm and an uncertainty-aware geometric encoder combined with DepthAnything3, whereas the semantic branch provides high-level semantic representations via a cross-view interaction module.1
• We further introduce a dual-branch balanced classifier-free guidance (CFG) strategy, which enhances the model's conditional generation capability while preserving the independent learning and collaborative controllability of the two branches.1
• Extensive experiments demonstrate that LiFi outperforms state-of-the-art methods in generating high-fidelity LiDAR scenes.1

Sources (1)

• [1]LiFi: LiDAR Generation from Multi-View Images via Geometric and Semantic Collaborative Guidance
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 08:34 AM

  Generating LiDAR point clouds from multi-view camera images is a valuable task with applications in controllable data synthesis, cross-modal simulation and various perception tasks.

  In this work, we introduce LiFi (LiDAR From images), a dedicated framework for generating visually and geometrically aligned LiDAR scenes from multi-view camera images.

Extractive summary: sentences quoted from the sources.