Scholar
Théo Sourget
Google Scholar ID: aidbxNUAAAAJ
PhD Student, PURRlab, IT University of Copenhagen
Citations & Impact
Citations
42
H-index
4
i10-index
2
Publications
7
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Contributed to multiple medical image analysis projects, including comparisons between Mask-RCNN and Detection Transformer, and classical data augmentation techniques versus new panoramic generation methods.
Research Experience
- PhD Student at PURRlab, IT University of Copenhagen (September 2025 - Current), studying data quality and the evaluation of machine learning algorithms.
- Research Engineer at MICS, CentraleSupélec (April 2025 - September 2025), studied biases in Vision-Language models.
- Research Assistant at PURRlab, IT University of Copenhagen (October 2023 - January 2025), evaluated the usage of non-relevant information by CNN models in chest X-ray images and studied how public medical image datasets are referenced in research papers.
- Assistant Lecturer at IT University of Copenhagen (January 2024 - July 2024), prepared and taught lectures to bachelor’s students on Git, decision trees, ensemble of classifiers, and common problems in ML projects.
- Teaching Assistant at IT University of Copenhagen (October 2023 - December 2023), assisted in practical work for the Data In the Wild Course.
- Data Scientist Intern at Capgemini Engineering - Medic@ (April 2023 - September 2023), worked on teeth detection, segmentation, and numbering in dental panoramic X-rays.
- Developer Intern at See-d (April 2021 - July 2021), developed a storage-related data analysis website and created a dashboard with Qlik Sense.
Education
- Master's degree in Data Science from Université de Rouen (2021-2023), specializing in projects focused on the classification and segmentation of medical images using deep learning models, training transformer models with small datasets, and the effects of data augmentation and transfer learning.
Background
- PhD Student with an interest in Medical Image Analysis. Passionate about Data Science and AI.
Miscellany
- Proficient in Python and its libraries such as PyTorch, Pandas, scikit-learn, and numpy; experience with Streamlit for tool creation; API development, especially with FastAPI; used Docker for easier deployment of models or tools; Linux as primary OS for 5 years; experienced with git-based development.
Co-authors