Segolene Martin homepage

Ségolène Martin

I am working as a postdoctoral researcher at TU Berlin (Math+ project), in the image analysis team headed by Prof. Gabriele Steidl and in collaboration with Prof. Hanno Gottschalk. My current research is at the intersection of optimization, optimal transport and generative learning, with applications to image processing and computer vision.

Prior to that, I was a PhD student in applied mathematics at CentraleSupélec, Université Paris-Saclay, in the CVN / OPIS Inria team. My PhD was under the supervision of Prof. Jean-Christophe Pesquet (CVN, CentraleSupélec) and Prof. Ismail Ben Ayed (ILLS, ETS Montréal). I defended my thesis Variational methods for large scale data problems in imaging on January 26th, 2024. You can find the manuscript here.

Research interests

Large-scale continuous optimization
Majoration-Minimization & proximal algorithms, Expectation-Maximization, convergence theory
Inverse problems for image processing
Image restoration and reconstruction, PSF estimation
Image classification
Clustering and few-shot optimization-based methods, unbalanced few-shot, transductive learning
Text-vision models
Zero and few-shot classification with CLIP
Image generation
Flow-based models, GANs

Education

PhD in Applied Mathematics, 2020-2024
CentraleSupélec, Université Paris-Saclay
Master in Mathematics, Vision and Learning (MVA), 2019-2020
École Normale Supérieure Paris-Saclay
Agrégation in Mathematics, 2018-2019
École Normale Supérieure Paris-Saclay
Bachelor's Degree in Fundamental Mathematics, 2016-2017
École Normale Supérieure Paris-Saclay

Email
Github
Google Scholar
CV

Publications

Publication preceeded with a * are the ones where I am a main author.

Journal articles

* J. Ajdenbaum, E. Chouzenoux, S. Martin, C. Lefort, J.-C. Pesquet, A novel variational approach for multiphoton microscopy image restoration: from PSF estimation to 3D deconvolution, Inverse Problems, 2024.
J.-B. Fest, T. Heikkilä, I. Loris, S. Martin, L. Ratti, S. Rebegoldi, G. Sarnighausen, On a fixed- point continuation method for a convex optimization problem, Advanced Techniques in Optimization for Machine learning and Imaging (ATOMI), 2023.
*E. Chouzenoux, S. Martin, J.-C. Pesquet, A local MM subspace method for solving constrained variational problems in image recovery, Journal of Mathematical Imaging and Vision, 2022.

Conference Proceedings

*S. Martin, A. Gagneux, P. Hagemann, G. Steidl, PnP-Flow: Plug-and-Play Image Restoration with Flow Matching., arXiv preprint arXiv:2410.02423.
*S. Martin, Y. Huang, F. Shakeri, J.-C. Pesquet, I. Ben Ayed, Transductive zero-shot and few-shot CLIP, 2024 IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR) as a highlight (<10 % of papers), 2024.
*A. Sadraoui, S. Martin, E. Barbot, A. Laurent-Bellue, J.-C. Pesquet, C. Guettier, A transductive few-shot learning approach for classification of liver cancer histopathology images, 2024 IEEE International Symposium on Biomedical Imaging (ISBI), 2024.
*S. Martin, M. Boudiaf, E. Chouzenoux, J.-C. Pesquet, I. Ben Ayed, Towards practical few-shot query sets : Transductive minimum description length inference, Neural Information Processing Systems (NeurIPS), 2022.
M. Kahanam, L. Le-Brusquet, S. Martin, J.-C. Pesquet, A Non-Convex Proximal Approach for Centroid-Based Classification, IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2022.
*S. Martin, E. Chouzenoux, J.-C. Pesquet, A Penalized Subspace Strategy for Solving Large-Scale Constrained Optimization Problems, IEEE 29th European Signal Processing Conference (EUSIPCO), 2021.

Talks

October 2024 - Unbalanced few-shot classification, DATAIA workshop on Mathematical foundations of artifical intelligence, Sorbonne Center for AI, Paris, France.
December 2022 - Towards Practical Few-Shot Query Sets : Transductive Minimum Description Length Inference, Seminar of the ILLS, Montreal, Canada.
July 2022 - Numerical restoration of multiphoton images, Seminar of the XLIM, Limoges, France.
May 2022 - Penalized methods for solving constrained variational problems in image recovery, Mini-Symposium: Variational Methods for Inverse Problems in Imaging, 10th International Conference Inverse Problems Modeling and Simulation.
March 2022 - A Penalized Subspace Strategy for Solving Large-Scale Constrained Optimization Problems, Mini-Symposium: Novel Perspectives in Optimization and Machine Learning for Imaging, SIAM Conference on Imaging Science

I am also involved in several teaching activities. I have been a teaching assistant for the following courses:

Refresher course on optimization, Master 2 MVA, ENS Paris-Saclay, 2020-2023 - course taught by J.-C. Pesquet (exercises, correction)
Optimization, Master 1, CentraleSupelec, 2023 - course taught by J.-C. Pesquet
Data analysis and statistics, Bachelor 1, IUT Orsay, 2022 - course taught by S. Marduel
Python, Bachelor 1, IUT Orsay, 2022 - course taught by S. Marduel
Signal processing, Bachelor 2, IUT Orsay, 2021 - course taught by F. Alberge
Data analysis and statistics, Bachelor 1, IUT Orsay, 2021 - course taught by T. Guilbaud
Python, Bachelor 1, IUT Orsay, 2021 - course taught by T. Guilbaud
Mathematics, Bachelor 1, IUT Orsay, 2020 - course taught by M. Lagron
Oral interrogations in mathematics in "Classes Préparatoires aux Grandes Ecoles", Lycée Saint Louis, Paris, 2019-2020