9 papers
DreamTraj: Generating 6-DoF Object Trajectories by Reading Unrendered Video Diffusion Latents
Tongsheng Ding, Zhen Luo, Yixuan Yang +4
Accurate prediction of object trajectories during manipulation is essential for closing the perception-action loop. Progress is limited on two fronts: available datasets lack fine-…
STABLE: Simulation-Ready Tabletop Layout Generation via a Semantics-Physics Dual System
Zhen Luo, Yixuan Yang, Xudong Xu +5
Generating simulation-ready tabletop scenes from task instructions is an intriguing and promising research direction in the field of Embodied AI. However, existing task-to-scene ge…
Code-as-Room: Generating 3D Rooms from Top-Down View Images via Agentic Code Synthesis
Yixuan Yang, Zhen Luo, Wanshui Gan +5
Designing realistic and functional 3D indoor rooms is essential for a wide range of applications, including interior design, virtual reality, gaming, and embodied AI. While recent…
Delta-K: Boosting Multi-Instance Generation via Cross-Attention Augmentation
Zitong Wang, Zijun Shen, Haohao Xu +2
While Diffusion Models excel in text-to-image synthesis, they often suffer from concept omission when synthesizing complex multi-instance scenes. Existing training-free methods att…
ArtiWorld: LLM-Driven Articulation of 3D Objects in Scenes
Yixuan Yang, Luyang Xie, Zhen Luo +4
Building interactive simulators and scalable robot-learning environments requires a large number of articulated assets. However, most existing 3D assets in simulation are rigid, an…
MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning
Jinkun Hao, Naifu Liang, Zhen Luo +8
The ability of robots to interpret human instructions and execute manipulation tasks necessitates the availability of task-relevant tabletop scenes for training. However, tradition…