SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . To address this, we propose SIMART, a unified MLLM framework that jointly performs part-level decomposition and kinematic prediction By introducing a Sparse 3D VQ-VAE, SIMART reduces token counts by 70% vs dense voxel tokens, enabling high-fidelity multi-part assemblies
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . To address this, we propose SIMART, a unified MLLM framework that performs part-level decomposition and kinematic prediction jointly By introducing a Sparse 3D VQ-VAE, SIMART reduces token counts by 70% vs dense voxel tokens, enabling high-fidelity multi-part assemblies
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . This work proposes SIMART, a unified MLLM framework that jointly performs part-level decomposition and kinematic prediction, and achieves state-of-the-art performance on PartNet-Mobility and in-the-wild AIGC datasets, and enables physics-based robotic simulation
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . To address this, we propose SIMART, a unified MLLM framework that jointly performs part-level decomposition and kinematic prediction By introducing a Sparse 3D VQ-VAE, SIMART reduces token counts by 70% vs dense voxel tokens, enabling high-fidelity multi-part assemblies
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . SIMART tackles this gap by teaching a single multimodal large model to do the two critical jobs at once: (1) break a monolithic mesh into meaningful parts (handle, door, frame), and (2) predict how those parts are connected and should move (the kinematics)
SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets . . . To address this, we propose SIMART, a unified MLLM framework that jointly performs part-level decomposition and kinematic prediction By introducing a Sparse 3D VQ-VAE, SIMART reduces token counts by 70% vs dense voxel tokens, enabling high-fidelity multi-part assemblies