The University of Hong Kong; University of Electronic Science and Technology of China (Xingjian Ran, Xiaoye Mo, Bo Dai et al.) · Image to 3D scenes
SceneMosaic
SceneMosaic turns a single indoor image into a complete, editable 3D scene. Its pipeline runs perception, promptable segmentation, per-object 3D reconstruction, inter-object relation extraction and scene evolution, then publishes representative scene variants as .glb files. It builds on the gated SAM3 and SAM 3D Objects models and needs an OpenAI-compatible API key for the perception, relation and evolution stages. It requires Linux and an NVIDIA GPU with at least 32 GB of VRAM (observed peak about 38 GB), and ships four example images. No licence is stated.
Open the source ↗Similar 3D & Spatial releases
- CoLa3D: Composable Latent 3D Decomposition
Xudong Cai, Yongcai Wang, Deying Li, Chuanxia Zheng (Renmin University of China, Hebei Key Laboratory RIAS, Nanyang Technological University) · Scene 3D decomposition · Week 46, 2026
- LucidLyra
NVIDIA Research (Oliver Hahn, Rodrigo Marcuzzi, Sherwin Bahmani, Jiahui Huang, Laura Leal-Taixé and others) · Queryable 3D scenes · Week 46, 2026
- StableGen
sakalond · Blender 3D texturing · Week 46, 2026
- PhysAg3D
Peking University, Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ), Huawei and Wuhan University · Interactive 3D agent · Week 45, 2026
- Idol
Speridlabs (Alejandro Cañada, Chema Garabito) · Image-to-3D splats · Week 45, 2026
- JST-2
Viggle AI · World Model · Week 44, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.