Integrated Texture and Material Synthesis
Extend geometry generation models with parallel texture prediction heads so output meshes include both shape and color.
Source paper
Tanghui Jia, Dongyu Yan, Dehao Hao, Yang Li, Kaiyi Zhang, Xianyi He, Lanjiong Li, Jinnan Chen, Lutao Jiang, Qishen Yin, Long Quan, Ying-Cong Chen, Li Yuan
Generated by AI. The paper’s authors did not write these.
Extend geometry generation models with parallel texture prediction heads so output meshes include both shape and color.
Replace marching cubes with a differentiable iso-surface layer so shape refinement produces clean, quad-dominant mesh topologies.
Adopt an unsigned distance field representation so shape generation models can accurately synthesize open and thin-shell surfaces.
This paper presents UltraShape 1.0, a scalable 3D diffusion framework designed to generate high-fidelity 3D geometry from single images. The authors introduce a two-stage pipeline where a coarse global structure is first synthesized using a vector set representation, followed by a voxel-based refinement stage that decouples spatial localization from detail synthesis to produce intricate, watertight meshes.
The method employs a robust data curation pipeline involving voxel-based watertightening and VLM-based filtering to create a high-quality dataset. For generation, it uses a cascaded Diffusion Transformer (DiT) architecture: Stage 1 generates a coarse shape using a vector-set DiT (based on Hunyuan3D-2.1), and Stage 2 refines this shape using a voxel-conditioned DiT with Rotary Positional Embeddings (RoPE) to synthesize high-frequency geometric details on a fixed grid.