Zhongang Cai

Staff Research Scientist@SenseTime Research·新加坡·OpenReview

研究方向

Spatial Intelligence · Multimodal Model · Digital Embodied Intelligence · Autonomous Characters · Human Motion Generation · Human Pose and Shape Estimation · Point Clouds

5.5

ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization

ICML 2025Poster

6.0

MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers

ICLR 2025Poster

5.0

U3D: Unlocking the Video Prior for High Fidelity Sparse Novel View Synthesis and 3D Generation

合作者 (20)

Zhongang Cai

The Quest for Generalizable Motion Generation: Data, Model, and Evaluation

Learning Text-driven 3D Human Motion Generation from 3D-free Web Videos

Controllable Human-centric Keyframe Interpolation with Generative Prior

ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization

MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformers

U3D: Unlocking the Video Prior for High Fidelity Sparse Novel View Synthesis and 3D Generation