Dongyang Jin  金冬阳

About Me

I am an AI researcher passionate about generative models and visual understanding. My research spans AIGC, image & video generation, and computer vision. If you find our work helpful, citations are greatly appreciated. Feel free to reach out via email or WeChat. Collaboration and discussion are always welcome.

Research Interests

News

• 2025.07 🎉🎉 One paper (SCALAR++) is accepted by IJCV.
• 2025.07 🎉🎉 One paper (FluxText) is accepted by TMM.
• 2026.06 🚀 Continue working for SI Team, Xiaohongshu as Algorithm Engineer.
• 2026.03 🚀 Joined SI Team, Xiaohongshu as Research Intern.
• 2026.02 🎉🎉 Two papers are accepted by CVPR 2026.
• 2025.11 🎉🎉 One paper is accepted by AAAI 2026.
• 2025.11 🏆 Awarded the National Scholarship (国家奖学金).
• 2025.06 🎉🎉 One paper is accepted by TPAMI.
• 2025.06 🎉🎉 One paper is accepted by MICCAI 2025.
• 2025.04 🚀 Joined DreamX Team, AMAP, Alibaba Group as Research Intern.
• 2025.03 🎉🎉 One paper is accepted by CVPR 2025.
• 2025.01 🚀 Joined YouTu Lab, Tencent as Research Intern.
• 2024.12 🎉🎉 One paper is accepted by AAAI 2025.
• 2023.12 🎉🎉 One paper is accepted by AAAI 2024.

Education

• 
  M.Eng. in Computer Science and Engineering

  Southern University of Science and Technology (SUSTech)

  Sep 2023 – Jul 2026
• 
  B.Eng. in Computer Science and Engineering

  Southern University of Science and Technology (SUSTech)

  Sep 2019 – Jul 2023

Experience

• 
  Research Intern & Algorithm Engineer · SI Team

  Xiaohongshu, Beijing

  Mar 2026 – Present
• 
  Research Intern · DreamX Team

  AMAP, Alibaba Group, Beijing

  Apr 2025 – Feb 2026
• 
  Research Intern · YouTu Lab

  Tencent, Shanghai

  Jan 2025 – Mar 2025

Publications

* Equal contribution  ·  † Corresponding author  ·  CCF A   CCF B   CCF C   Preprint  ·  Total citations: 347

• 
  SCALAR++: Efficient Controllable Generation via Scale-wise Visual Autoregressive Learning

  R. Xu*, D. Jin*, S. Chen*, Y. Bai, J. Ma, R. Lan, J. Zeng, Y. Ge, M. Zheng, L. Sun, X. Chu

  CCF A IJCV, 2026
• 
  Semantic Context Matters: Improving Conditioning for Autoregressive Models

  D. Jin*, R. Xu*, J. Zeng, R. Lan, Y. Bai, L. Sun, X. Chu

  CCF A CVPR 2026
• 
  Eevee: Towards Close-up High-Resolution Video-based Virtual Try-on

  J. Zeng, Y. Bai, R. Chen, X. Zhang, L. Sun, D. Jin, R. Xu, N. Zhang, D. Song, et al.

  CCF A CVPR 2026 Findings
• 
  SCALAR: Scale-wise Controllable Visual Autoregressive Learning

  R. Xu*, D. Jin*, Y. Bai, R. Lan, X. Duan, L. Sun, X. Chu

  CCF A AAAI 2026
• 
  SAFE: A Semantic-Appearance Full-Body Editor for Pedestrian Video Anonymization

  J. Ma, C. Fan, D. Ye, J. Yang, D. Jin, C. Luo, J. Li, F.M. Hassan, S. Yu

  Preprint Under Review, 2026
• 
  On Denoising Walking Videos for Gait Recognition

  D. Jin*, C. Fan*, J. Ma, J. Zhou, W. Chen, S. Yu

  CCF A CVPR 2025
• 
  Exploring More from Multiple Gait Modalities for Human Identification

  D. Jin*, C. Fan*, W. Chen, S. Yu

  CCF A AAAI 2025
• 
  OpenGait: A Comprehensive Benchmark Study for Gait Recognition towards Better Practicality

  C. Fan, S. Hou, J. Liang, C. Shen, J. Ma, D. Jin, Y. Huang, S. Yu

  CCF A TPAMI 2025
• 
  Flux-text: A Simple and Advanced Diffusion Transformer Baseline for Scene Text Editing

  R. Lan, Y. Bai, X. Duan, M. Li, D. Jin, R. Xu, D. Nie, L. Sun, X. Chu

  CCF A TMM, 2026
• 
  Pose as Clinical Prior: Learning Dual Representations for Scoliosis Screening

  Z. Zhou, Z. Peng, D. Jin, C. Fan, F. An, S. Yu

  CCF B MICCAI 2025
• 
  SkeletonGait: Gait Recognition using Skeleton Maps

  C. Fan, J. Ma, D. Jin, C. Shen, S. Yu

  CCF A AAAI 2024
• 
  Passersby-Anonymizer: Safeguard the Privacy of Passersby in Social Videos

  J. Ma, H. Luo, Z. Huang, D. Jin, R. Wang, J.A. Briffa, N. Poh, S. Yu

  CCF C IJCB 2024

Projects

• 
  FLUX-Text: Scene Text Editing via Diffusion Transformer

  An open-source scene text editing project supporting multi-line text in complex visual scenes (posters, ads, games) for both English and Chinese, built on FLUX with LoRA fine-tuning.

  Code ★ 777 Stars
• 
  OpenGait: A Flexible and Extensible Framework for Gait Recognition

  An open-source gait recognition framework supporting multiple SOTA models and datasets (CASIA-B, GREW, Gait3D, SUSTech1K, etc.), with DDP training and AMP support. CVPR 2023 highlight paper, extended to TPAMI 2025.

  Code ★ 1.1k Stars

Services

• Teaching Assistant

• 
  Python Programming

  Southern University of Science and Technology (SUSTech)

  Feb 2023 – Jun 2024
• 
  Software Engineering

  Southern University of Science and Technology (SUSTech)

  Sep 2022 – Feb 2023

• Conference Reviewer

CVPR  ·  ECCV  ·  AAAI

Honors & Awards

• Nov 2025 National Scholarship（国家奖学金）