ZJU | CHINGMU Advances Scalability of Controllable Human Animation Synthesis_Entertainment-SHANGHAI CHINGMU VISION TECHNOLOGY CO., LTD

2026-06-10

ZJU | CHINGMU Advances Scalability of Controllable Human Animation Synthesis

Zhejiang University，ZJU

Computer Graphics

Data Acquisition Facility（DAQ）

Human Animation Synthesis

Professor Weiwei Xu's team at Zhejiang University's State Key Laboratory of CAD&CG published "Neural Motion Graphs" at SIGGRAPH Asia, the top computer graphics conference. The research used CHINGMU's system to collect high-quality human motion data, building an HMD dataset covering 17 movement types including walking, running, gymnastics and dance.

The team proposed a novel framework combining motion graphs with neural networks, implementing highly controllable, scalable real-time motion synthesis via "neural nodes" and "neural edges". Unlike traditional methods, neural motion graphs can flexibly add new movement types without degrading existing generation quality and enable natural transitions between movements, opening new directions for motion capture and computer graphics integration.

Research Achievements：https://dl.acm.org/doi/10.1145/3610548.3618181

Obtain exclusive case studies and industry solutions

Get the complete case in one click. Our experts will precisely match the most suitable motion capture solution for your industry, helping you reduce costs and increase efficiency.

ZJU | CHINGMU Advances Scalability of Controllable Human Animation Synthesis

Obtain exclusive case studies and industry solutions

Name*

Email*

Country*

Enterprise / Institution*

Application fields*

Please leave a message.*