Qingyuan Cai

Qingyuan Cai

Beijing Normal University

Flag Counter

About Me

👋 Hello, I am Qingyuan Cai. I am a second year Ph.D. student at Beijing Normal University. I was fortunate to be mentored by Yongzhen Huang and Saihui Hou. I’m also currently working as an intern at ROBOTERA, focusing on Egocentric Videos Pretraining for VLA models.

More broadly, my research spans Human Pose Estimation, Embodied AI, Video Understanding and Biometrics.

🔥 News

  • [Mar. 2026] 🎉 One paper accepted by ICME 2026.
  • [Feb. 2026] 🎉🎉 Two papers accepted by CVPR 2026.
  • [Dec. 2025] A Unified Open-source 3D Human Pose Estimation framework released on arXiv.
  • [Feb. 2025] 🎉 One paper accepted by AAAI 2025.
  • [Oct. 2024] 🎉 One paper accepted by ACM MM 2024.
  • [Feb. 2024] 🎉 One paper accepted by AAAI 2024.
  • [Sep. 2023] 🎉 One paper accepted by TMM.

📚 Publications and Preprints

BarbieGait: An Identity-Consistent Synthetic Human Dataset with Versatile Cloth-Changing for Gait Recognition
CVPR

BarbieGait: An Identity-Consistent Synthetic Human Dataset with Versatile Cloth-Changing for Gait Recognition

Qingyuan Cai, Saihui Hou, Xuecai Hu, Yongzhen Huang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.

Paper Code Project Page BibTex
MMGait: Towards Multi-Modal Gait Recognition
CVPR

MMGait: Towards Multi-Modal Gait Recognition

Chenye Wang†, Qingyuan Cai†, Saihui Hou, Aoqi Li, Yongzhen Huang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.

Paper Code BibTex


† Equal contribution

FastDDHPose: Towards Unified, Efficient, and Disentangled 3D Human Pose Estimation
arXiv

FastDDHPose: Towards Unified, Efficient, and Disentangled 3D Human Pose Estimation

Qingyuan Cai, Linxin Zhang, Xuecai Hu, Saihui Hou, Yongzhen Huang

arXiv 2512.14162, Preprint, 2025.

Paper Code BibTex
Disentangled Diffusion-Based 3D Human Pose Estimation with Hierarchical Spatial and Temporal Denoiser
AAAI

Disentangled Diffusion-Based 3D Human Pose Estimation with Hierarchical Spatial and Temporal Denoiser

Qingyuan Cai, Xuecai Hu, Saihui Hou, Li Yao, Yongzhen Huang

Proceedings of the AAAI conference on artificial intelligence (AAAI), 2024.

Paper Code BibTex
Training-Free High-Resolution Image Generation via Optimized Attention Computation
ICME

Training-Free High-Resolution Image Generation via Optimized Attention Computation

Junjie Li, Zengbin Wang, Qingyuan Cai, Yafeng Chen, Man Zhang

2026 IEEE International Conference on Multimedia and Expo (ICME 2026).

Gait recognition with drones: A benchmark
TMM

Gait recognition with drones: A benchmark

Aoqi Li, Saihui Hou, Qingyuan Cai, Yang Fu, Yongzhen Huang

IEEE Transactions on Multimedia (TMM), 2023.

Paper Code BibTex
AerialGait: Bridging Aerial and Ground Views for Gait Recognition
ACM MM

AerialGait: Bridging Aerial and Ground Views for Gait Recognition

Aoqi Li, Saihui Hou, Chenye Wang, Qingyuan Cai, Yongzhen Huang

Proceedings of the 32nd ACM International Conference on Multimedia (ACM MM), 2024.

Paper BibTex
RA-GAR: A Richly Annotated Benchmark for Gait Attribute Recognition
AAAI

RA-GAR: A Richly Annotated Benchmark for Gait Attribute Recognition

Chenye Wang, Saihui Hou, Aoqi Li, Qingyuan Cai, Yongzhen Huang

Proceedings of the AAAI conference on artificial intelligence (AAAI), 2025.

Paper BibTex

🏅 Awards

  • 🏆 2024.10, National Scholarship for Graduate Students.
  • 🥉 2024.10, 3rd Place (Pose) at ACM MM’24 Gait Challenge.

🤝 Services

Conference Reviewer