AI Paper Atlas

Research map / 3D vision / Depth and stereo

VGGT-Ω

Jianyuan Wang, Minghao Chen, Shangzhan Zhang, Nikita Karaev, Johannes Schönberger, Patrick Labatut, Piotr Bojanowski, David Novotny, Andrea Vedaldi, Christian Rupprecht

CVPR 2026 · 3D vision › Depth and stereo · approach: Feed-forward 3D reconstruction · 97 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
scale training data from 2K to 2M sequences (ablation)VGGT-Omega 1B · point error (1B ablation)point error: 0.275 -> 0.073Improved
scale model from 1B to 10B (ablation)VGGT-Omega · Sintel camera poseAUC@3: 35.5 -> 40.0Improved
replace 25% of global attention layers with register attention (ablation)VGGT-Omega 1B · point error (1B ablation)point error: 0.071 -> 0.073No clear effect
replace 25% of global attention with register attention (speed) (ablation)VGGT-Omega · inference speedspeedup: 20-25% speedupImproved
replace all global attention with register attention (ablation)VGGT-Omega · 1000-frame inferenceruntime seconds: 240.2 -> 11.7Improved
add point and matching losses to multi-task training (ablation)VGGT-Omega 1B · point error (1B ablation)point error: 0.078 -> 0.073Improved
use VGGT's multi-head multi-task setup instead of single dense head (ablation)VGGT-Omega 1B · point error (1B ablation)point error: 0.073 -> 0.070No clear effect
add 10% self-supervised training steps (ablation)VGGT-Omega 1B · point error (1B ablation)point error: 0.073 -> 0.070Improved
use own annotation pipeline instead of MegaSaM pseudo labelspseudo ground truth · Sintel annotation qualityAUC@30 camera: 62.1 -> 96.4Improved
add frozen VGGT-Omega scene tokens to OpenVLA-OFTOpenVLA-OFT · LIBEROaverage success rate: 97.1 -> 98.5Improved
use VGGT-Omega 10B over DA3camera pose · SintelAUC@3: 16.2 -> 40.0Improved
use text-only LLM embedding instead of VLM embedding for retrievallanguage-aligned registers · language retrievaltop-1 accuracy: 76.8 -> 47.5Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy