Latent Bridge: From Video Latents to 3D Geometry via Feature Alignment

Key points

• Video generation and feed-forward 3D reconstruction typically use separate representations, requiring RGB decoding and visual re-encoding to recover geometry from generated latents.1
• We develop representation-specific bridges for Wan2.1 and V-RAE, enabling feed-forward prediction of cameras, depth, and 3D point clouds without the intermediate RGB pathway.1
• At inference, the same video latents support both RGB decoding and geometry prediction.1
• Experiments on Co3Dv2 and WildRGB-D demonstrate competitive geometric reconstruction quality, efficient end-to-end generation and reconstruction, and faster latent-to-geometry decoding, highlighting latent alignment as a practical and reusable interface between video generation and 3D reconstruction.1

Sources (1)

• [1]Latent Bridge: From Video Latents to 3D Geometry via Feature Alignment
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 04:57 PM

  We propose Latent Bridge, which connects video generative latent spaces with geometric feature spaces to directly predict 3D geometry from video latents.

  Video generation and feed-forward 3D reconstruction typically use separate representations, requiring RGB decoding and visual re-encoding to recover geometry from generated latents.

Extractive summary: sentences quoted from the sources.