Mahshid Sadeghpour

Mahshid Sadeghpour

Cyber Security Researcher

I am a Lecturer at the School of Science at RMIT University, holding a Ph.D. in Applied Mathematics. My research interests include Cyber Mathematics, Biometrics, Privacy-Preserving Machine Learning, and Children’s Online Safety.

I have developed the first privacy-preserving retinal recognition system and created the first large-scale high-quality synthetic retinal image dataset using data-driven generative modelling. My contributions have earned numerous awards for research excellence, innovative teaching, and community outreach.

I am passionate about research, teaching, and industry collaboration. Outside of work, I enjoy coffee, ice-skating, and learning Chinese.

Open Research Projects

  • Generative modelling for synthesising biometrics images
  • Semi-supervised and self-supervised learning for biometrics system development
  • Machine unlearning for regulatable biometrics
  • Privacy protection technologies for children’s e-safety

Enquiries from master and Ph.D. applicants are welcome. 欢迎有意申请硕博项目的同学来信咨询。

Research Grants

  • Sharenting Practices on Instagram in Australia: Threats and Countermeasures. Boundary Crossers Research Grant. RMIT STEM College, 2024
  • Diffusion Modelling and Reinforcement Learning to Overcome Bias in Synthetic Biometric Data. Seed Funding. RMIT Center for Cyber Security Research and Innovation, 2024

Awards

  • Cyber Security Educator of the Year Award (Master of Cyber Security Teaching Team). Australian Information Security Association, 2024
  • RMIT Vice-Chancellor’s Teaching Award for Programs and Services that Enhance Learning. RMIT University, 2024
  • School of Science Outreach Award. RMIT University, 2024
  • HDR Science Prize. RMIT University, 2024
  • RMIT HDR Prize for Research Excellence. RMIT University, 2023
  • STEM Award for Initiatives that Enhance Learning (Master of Cyber Security Teaching Team). RMIT University, 2023
  • The Young Scientist Research Prize. The Royal Society of Victoria, 2022

Publications

  • Generative AI in Collaborative Academic Report Writing: Advantages, Disadvantages, and Ethical Considerations. M Sadeghpour, A Arakala, A Rao. arXiv preprint arXiv:2504.08832, 2025
  • Protection of sparse retinal templates using cohort-based dissimilarity vectors. M Sadeghpour, A Arakala, SA Davis, KJ Horadam. IEEE Transactions on Biometrics, Behavior, and Identity Science 5 (2), 233-243, 2023
  • GRETINA: A large-scale high-quality generated retinal image dataset for security and privacy assessment. M Sadeghpour, A Arakala, SA Davis, KJ Horadam. International Conference on Image and Vision Computing New Zealand, 373-387, 2022
  • Failure of affine‐based reconstruction attack in regenerating vascular feature points. M Sadeghpour, A Arakala, SA Davis, KJ Horadam. IET Biometrics 10 (5), 497-517, 2021
  • Application of affine-based reconstruction to retinal point patterns. M Sadeghpour, A Arakala, SA Davis, KJ Horadam. International Conference of the Biometrics Special Interest Group, 2020