AI
VGGT-Diff: Visual Geometry Meets Diffusion for Sparse-View Novel View Synthesis
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 26/09/2026, 20:00:00 UTC
Briefing created: 29/09/2026, 04:24:49 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 29/09/2026, 04:18:25 UTC.
Briefings
VGGT-Diff is a geometry-routed multi-view diffusion model designed for sparse-view novel view synthesis. It bridges reconstruction-based and diffusion-based approaches by routing visual geometry latents from VGGT-\(\Omega$ into a pretrained video diffusion model. Each visual token is associated with a 3D point and confidence, transformed into query-aligned latent conditions via a confidence-aware Visual Geometry Router (VGR). These conditions guide joint target-view denoising, while Point-Track Residual Consistency (PTRC) regularizes predicted-clean residuals along reliable 3D tracks. Experiments demonstrate competitive or state-of-the-art performance across interpolation and extrapolation under various viewpoint difficulties.
Key points
- VGGT-Diff uses a confidence-aware Visual Geometry Router (VGR) to transform visual tokens into query-aligned latent conditions that guide joint target-view denoising.
- Point-Track Residual Consistency (PTRC) regularizes predicted-clean residuals along reliable 3D tracks to improve multi-view stability during inference.
Context and limitations
Performance is evaluated under different viewpoint difficulties, but specific quantitative metrics for each difficulty level are not provided in the evidence.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 29/09/2026, 04:24:49 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeif6libwo2tfa7fr7dl2a6pph3iwxcivpouq67tpvg3hcqahmkh62y