Quanta AI Labs

ByteDance Seed · Real-to-Sim Scenes

Lucida

Composable real-to-sim scene modelling rebuilding an indoor scene as complete, editable object assets rather than one fused mesh, through a parse-generate-place pipeline: parse per-instance evidence from posed RGB, depth, mask and point-cloud observations, generate occlusion-free object images and 3D assets, then place them with GizmoAct, a vision-language model that closed-loop edits object poses inside a 3D editor. Evaluated on R2S-Scene, CA-1M and Aria Digital Twin for detection, pose and Chamfer distance against Boxer, WildDet3D, Any6D, FoundationPose and SAM 3D.

Open the source ↗

Similar 3D & Spatial releases

  • Idol

    Speridlabs (Alejandro Cañada, Chema Garabito) · Image-to-3D splats · Week 45, 2026

  • JST-2

    Viggle AI · World Model · Week 44, 2026

  • UniForge3D

    Peking University, Math Magic and HKUST · Image-to-3D · Week 43, 2026

  • GenIA

    Meta Reality Labs and University of Tübingen · Input-Aligned 3D Reconstruction · Week 42, 2026

  • InSpatio-World 1.5

    InSpatio · Explorable World Model · Week 41, 2026

  • FIRE3D

    UIUC & Cornell · Minute-Scale Reconstruction · Week 37, 2026

More from ByteDance Seed

Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.