Stelios Perrakis

prof_pic2.png

I was born and raised in Chania, Crete, where I became interested early in science, math, and problem-solving. Growing up in a place that values independence and resilience shaped how I work today, with discipline, curiosity, and responsibility for what I do.

I completed my MSc in Computer Engineering at the University of Crete (GPA: 9.14/10), where I specialized in computer vision, neural networks, and pattern recognition. My master’s thesis focused on generative models for super-resolution in medical imaging, which I developed during my time as a Research Fellow at the Foundation for Research & Technology Hellas (FORTH).

I am currently a PhD candidate at CentraleSupélec, University of Paris-Saclay, where I work on deep learning methods for digital pathology. My research focuses on developing techniques for large medical image datasets under weak supervision, in settings where labels are limited, data are noisy, and decisions matter. I am particularly interested in generative models, representation learning, and their applications to real-world clinical problems.


Keywords: Digital Pathology, Generative Models, Deep Learning, Computer Vision

news

Oct 06, 2026 New preprint: Frame the adversary: a structure-aware attack methodology, with Vicky Kouni, Francis Bach, Pascal Frossard and Yann Chevaleyre, is now on arXiv.

selected publications

  1. Frame the adversary: a structure-aware attack methodology
    Vicky Kouni, Stelios Perrakis , Francis Bach, and 2 more authors
    2026
  2. ML-Based Multi-Horizon Wind Speed and Wind Direction Forecasting for Aviation and Energy Applications in Coastal Crete
    K. Blazakis, S. Perrakis, and  others
    Electronics, Dec 2025