CV

Haoxin Yang

South China University of Technology

📧 harxis@outlook.com


🎓 Education

South China University of Technology, Guangzhou, China

  • Ph.D. in Computer Science and Technology
  • Sep. 2022 – Jun. 2026 (Expected)
  • Advisor: Prof. Xuemiao Xu
  • Research: Privacy and security in generative models

Shenzhen University, Shenzhen, China

  • M.Eng. in Electronic and Communication Engineering
  • Sep. 2019 – Jun. 2022
  • Advisors: Prof. Bin Li, Prof. Yi Wang
  • Research: Privacy protection in federated learning

South China Agricultural University, Guangzhou, China

  • B.Eng. in Electronic Information Science and Technology
  • Sep. 2015 – Jun. 2019

📄 Publications

*: Co-first author †: Corresponding author 8 papers in CCF-A / CAS Q1 venues, including 3 first-author papers

  1. StableGuard: Towards Unified Copyright Protection and Tamper Localization in Latent Diffusion Models NeurIPS 2025 (CCF-A)

  2. StarPose: 3D Human Pose Estimation via Spatial-Temporal Autoregressive Diffusion IEEE TCSVT 2025 (CCF-B, CAS Q1 TOP)

  3. G2Face: High-Fidelity Reversible Face Anonymization via Generative and Geometric Priors IEEE TIFS 2024 (CCF-A, CAS Q1 TOP)

  4. Individual Property Inference Over Collaborative Learning in Deep Feature Space ICME 2022 (CCF-B)

  5. Beyond Inference Intervention: Identity-Decoupled Diffusion for Face Anonymization IEEE TIFS (Under Review, 2025)

  6. ViewSRD: 3D Visual Grounding via Structured Multi-View Decomposition ICCV 2025 (CCF-A)

  7. SITA: Structurally Imperceptible and Transferable Adversarial Attacks for Stylized Image Generation IEEE TIFS 2025 (CCF-A)

  8. EntropicSurge: Chaotic Latents are Principled Image-to-Video Attackers IEEE TIFS (Under Review, 2025)

  9. SCJD: Sparse Correlation and Joint Distillation for Efficient 3D Human Pose Estimation ICME 2025

  10. GenPoly: Learning Generalized and Tessellated Shape Priors via 3D Polymorphic Evolving IEEE TPAMI 2025 (CCF-A, CAS Q1)

  11. Hierarchical Damage Correlations for Old Photo Restoration Information Fusion 2025 (CAS Q1)

  12. Registration is a Powerful Rotation-Invariance Learner for 3D Anomaly Detection NeurIPS 2025 (CCF-A)


🧠 Patents

  1. Dynamic Visual Intelligent Vending Cabinet Product Detection, Tracking and Post-processing Method CN118298356A (Under Examination)

  2. Data Reconstruction Method and System Based on Deep Features and Gradient Information CN114330514B (Granted)

  3. Method and System for Inferring Data Attributes in Collaborative Learning CN114287009B (Granted)


🚀 Research & Project Experience

Large-Scale Urban Traffic Operation Safety (National Key R&D Program)

  • Subproject Leader

  • Jan. 2023 – Aug. 2025

  • Focus: Privacy protection for large-scale traffic visual data

  • Contributions:

    • GAN + geometric prior-based face anonymization
    • Diffusion-based anonymization framework
  • Result published in IEEE TIFS


Copyright Protection & Tamper Localization in Latent Diffusion Models

  • Project Lead

  • Jun. 2024 – Sep. 2025

  • Proposed StableGuard:

    • Unified framework for copyright protection and tamper detection
    • Self-supervised watermark embedding via multiplexed VAE
    • Mixture-of-experts forensic network
  • Published in NeurIPS 2025


Monocular Video-based 3D Human Pose Estimation

  • Project Lead

  • Jun. 2024 – Aug. 2025

  • Contributions:

    • StarPose: diffusion-based autoregressive framework
    • SCJD: sparse correlation + joint distillation

Intelligent Unmanned Retail System

  • Technical Lead
  • Apr. 2023 – Jun. 2025
  • Developed YOLO-based detection & tracking system
  • Successfully deployed in industry

🧪 Academic Services

  • Reviewer for: IEEE TIFS, TCSVT, TAI, ICME, IMAVIS

🏆 Honors & Awards

  • First Prize, Greater Bay Area Graduate Academic Forum, 2025
  • Graduate Academic Scholarship, Shenzhen University, 2020

🛠 Skills

Research Skills

  • Independent research, data analysis, academic writing, presentation

Technical Skills

  • Python, PyTorch, LaTeX, Shell, Linux, Docker

Languages

  • Chinese (Native), English (Fluent), Cantonese, Teochew