Search results

Abstract Deep learning has revolutionized computer vision in recent years and has been applied to many fields. This thesis focuses on medical endoscopy, where deep learning can assist physicians in many tasks, such as navigating the lungs during bronchoscopy, assisting in the detection of lung diseases, detecting Crohn's disease from capsule endoscopy (PillCam), or automating the detection of polyps during colonoscopy procedures.This thesis, entitled From Pixels to Practice: Deep Learning for Endoscopy, explores how modern neural networks and learning paradigms can improve visual understanding in endoscopy, with the aim of contributing to computer-aided detection (CAD) systems that can be integrated into clinical workflows.This work follows an article-based structure and links methodological advances in geometric and temporal modeling to techniques for handling data scarcity and imbalance, as well as to the practical and clinical implications of deep learning for lung tumor detection, both from a clinical and practitioner perspective. The first part of the manuscript provides a common foundation for all subsequent parts. First, we present a general introduction to the field of machine learning in Chapter 1, explaining concepts such as classification, loss functions, and artificial neural networks. Next, Chapter 2 focuses on the field of deep learning for computer vision, detailing the main vision tasks, the concept of convolutional neural networks, ResNet, and U-Net. Finally, Chapter 3 describes medical imaging, with a focus on computed tomography (CT) scans and optical imaging. The second part of the thesis focuses on learning spatio-temporal representations. In Chapter 4, we use deep neural networks combining spatial features and temporal recurrence to address the problem of detecting the bronchial carina, an anatomical landmark that helps doctors navigate the lungs. By evaluating classification (ResNet-50), segmentation (nnU-Net), and recurrent (GRU) models on a bronchoscopy dataset we created, the study highlights the benefits of combining information from segmentation masks and temporal features. Chapter 5 continues the segmentation task by analyzing the extent to which rotation-equivariant U-Nets, based on E(2)-CNNs with C4, C8, and D4 symmetry groups, can improve performance when the orientation of objects in the image is arbitrary. Together, these chapters show how temporal and geometric modeling capture complementary aspects of visual structure. They further highlight that data imbalance and scarcity are recurring problems in deep learning. The third part studies learning in situations of data scarcity and imbalance. First, Chapter 6 explores supervised contrastive pre-training [1] on large, domain-close endoscopic datasets (Hyper-Kvasir [2], LDPolyp [3]), which is then transferred to smaller, disease-specific data (Crohn-IPI [4]). This methodology performs better than pre-training on ImageNet or based on cross-entropy, highlighting the value of domain-specific contrastive representations. Next, Chapter 7 introduces Mask-Aware Cropping (MAC), a new data augmentation technique that mitigates pixel-level imbalance in segmentation. On various datasets with varying imbalance regimes (URDE [5], Kvasir-SEG [6], HAM10000 [7]), MAC consistently improves Dice and IoU metrics under conditions of extreme imbalance. Together, these methods form a data-centric framework for effective learning when annotations are scarce or unevenly distributed. The fourth part of the thesis focuses on deep learning in the operating room. Chapter 8 proposes a first model (ResNet-50) for the visual detection of lung cancer in bronchoscopy, trained on real, in-vivo data. The model outperforms junior physicians, while remaining inferior to experts. This result shows that CAD systems for lung cancer detection are promising. Chapter 9 extends this work by evaluating the usability of a CAD system based on a deep learning model. Combining probability indices, temporal graphs, and saliency map overlays, a multicenter evaluation with 10 physicians is conducted. The tool received favorable feedback, with high usability (SUS score of 80.5 [8]) and strong clinical acceptance. Beyond endoscopy, the results concerning rotation equivariance and pixel imbalance can be generalized to other fields such as microscopy, dermatology, and aerial imaging. This shows that the proposed methods are applicable to visual learning under structured variability and limited data constraints.Keywords: machine learning, computer vision, medicine, endoscopy, convolutional neural networks, segmentation, recurrent models, equivariance.  Jury Prof. Bruno Dumas - University of NamurProf. Frénay Benoit - University of NamurProf. Schobbens P-Y. - University of NamurProf. Beuls Katrien - University of Namur,Dr. Benjamin Mertens - Lys MédicalProf. Oramas Mogrojevo José Antonio - University of AntwerpDr. Mancas Matei - University of Mons Register

Save the date! More information coming soon.

May 5, 2026 | Registration for Thibaut Septon’s public thesis defense Name First name E-mail address Will attend the reception following the defense Yes ( optional ) No ( optional ) Need a parking sticker Yes ( optional ) No ( optional ) Would like a certificate for defense assistance Yes ( optional ) No ( optional ) In order to process your request, you must complete all fields marked "optional". When you submit this form, the completed data will be transmitted to UNamur and used to process your request. Learn more about your data protection and your rights This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

At a time when misinformation, post-truths and conspiracies are undermining confidence in science, UNamur welcomes SPiN (Science & Philosophy in Namur), a new interdisciplinary research center that questions the place of science in society. Founded last September by Olivier Sartenaer, Professor of Philosophy of Science at UNamur, SPiN brings together philosophers and scientists around a common vision: to develop a critical and accessible reflection on science in all its diversity..

Conceptual field theory and didactic situation theory Two complementary theories for thinking about and organizing mathematical learningIn the first part of this talk, I'll propose an introductory situation in literal arithmetic (Barallobres & Giroux, 2008) that will enable me to introduce and illustrate the main concepts and methodological tools of the theory of didactic situations developed by Guy Brousseau (Brousseau, 1997; Bessot, 2024). In the second part, I will give a brief presentation of Gérard Vergnaud's theory of conceptual fields (Vergnaud, 1990; Durand-Guerrier and Saby, 2023), then show how this theory helps to shed further light on the introductory situation.The contributions of the anthropological theory of didactics The Anthropological Theory of Didactics (TAD, Chevallard, 2001) aims to explain why and how a given knowledge lives in a certain institution, and/or is transformed as it passes from one institution to another. This perspective and some of its evolutions will be presented and illustrated in this talk.Students' activities and teachers' practices in the mathematics classroom: analysis methodology with Activity TheoryIn this talk, we will present the founding assumptions of the Activity Theory framework adapted to the Didactics of Mathematics (TADM, Vandebrouck, 2008), showing how this theory gives importance to fine-grained analyses of mathematical knowledge to appreciate classroom developments. We will then exemplify some of these aspects on the teaching of limits at university (Bridoux and Grenier-Boley, 2024).BibliographyBarallobres, G., & Giroux, J. (2008). Environmental deficiencies and regulations in validation situations. N. In Berdnaz, & C. Mary (Eds). L'enseignement des mathématiques face aux défis de l'école et des communautés. Actes du colloque EMF 2006 (CD-ROM). Éditions du CRP https://emf.unige.ch/application/files/1414/5390/4857/EMF2006_GT8_Barallobres.pdfBessot, A. (2003). An introduction to the theory of didactic situations. Cahiers du laboratoire Leibniz, 91. hal-00078794Bridoux, S., & Grenier-Boley, N. (2024). What teaching practices should be used to introduce the limits of functions in the first year of university? A case study. In A. González-Martín, G. Gueudet, I. Florensa & N. Lombard (Eds.), Proceedings of the Fifth Conference of the International Network for Didactic Research in University Mathematics (INDRUM 2024, 10-14 June 2024) (pp. 791-800). Escola Universitària de Sarrià. Univ. Autònoma de Barcelona and INDRUM.Brousseau, G. (1997). Théorie des situations didactiques. Lecture given at the award to Guy Brousseau of the title of Doctor Honoris Causa from the University of Montreal. http://www.cfem.asso.fr/actualites/archives/Brousseau.pdfChevallard, Y. (2001). Organizing study: 1. Structures and Functions. In J.-L. Dorier, M. Artaud, M. Artigue, R. Berthelot, & R. Floris Proceedings of the XIe École d'été de didactique des mathématiques. (pp. 3-32). Editions la Pensée Sauvage.Durand-Guerrier, V., & Nicolas Saby, N. (2023). Usages de la théorie des champs conceptuels en didactique des mathématiques. The example of transitivity. Caminhos da Educação Matemática em Revista, 13 (4),118-134. ⟨hal-04585866⟩Vandebrouck, F. (dir.) (2008). La classe de mathématiques: activités des élèves et pratiques des enseignants. OCTARES Éditions.Vergnaud, G. (1990). Conceptual field theory. Recherches en didactique des mathématiques, 10(2/3), 133-170. In practice Program18:00: Presentation by Viviane Durand-Guerrier 18:40: Presentation by Ghislaine Gueudet 19:20: Presentation by Nicolas Grenier-Boley and Stéphanie Bridoux20:00: Q&A around the three presentationsModalities Required registration via the PhD school website or direct link to the registration form.Or online participation link (via TEAMS).

JuryProf. Guillaume BERIONNI (UNamur), PresidentProf. Stéphane VINCENT (UNamur), SecretaryProf. Carmen GALAN (University of Bristol)Dr. Louis FENSTERBANK (Collège de France)Prof. Raphaël ROBIETTE (Université catholique de Louvain)AbstractCarboxylic acids are ubiquitous in nature and inexpensive compounds. Decarboxylation has become a key chemical transformation and has been widely reported in organic chemistry except for carbohydrates. This reaction can be catalyzed by transition metal and can also be induced by light, thermal activation, or photocatalysis. Borylated compounds have stimulated the pharmaceutical industry's interest (Boromycin, Bortezomib or boron neutron capture therapy). Recent methodologies have been developed to transform carboxylic acids to boronate esters by metal-catalyzed or light-promoted or photocatalyzed reactions. In this thesis, we explored the synthesis of borylated carbohydrates through a decarboxylation pathway. More specifically, sialic acids being among the most important carbohydrates in glycobiology, we addressed the problem of the synthesis of borylated sialic acids. On the other hand, organophosphates play an important role in diverse fields: in materials chemistry, in agriculture, in organic chemistry, and in biochemistry. Phosphorylation is a key reaction in biological processes such as signal transduction and cell activity regulation. The formation of phosphorylated carbohydrates has been widely described through two-electron mechanisms. However, radical phosphorylation of carbohydrates remains unexplored. This Ph.D. thesis describes the development of new methodologies for the decarboxylative functionalization of carbohydrates, focusing on borylation and phosphorylation..

Our keynote speakers for 2026 are Professor Roosmarijn Vandenbroucke (Ghent University) and Professor Nelly Litvak (Eindhoven University of Technology). More information on the "Women in Science" website

The F.R.S.-FNRS has just published the results of its various 2025 calls for proposals. These include the "Credits & Projects" and "WelCHANGE" calls, as well as the "FRIA" (Fund for Research Training in Industry and Agriculture) and "FRESH" (Fund for Research in the Humanities) calls, which aim to support doctoral theses. What are the results for UNamur? Twenty-eight projects have been selected, demonstrating the quality and richness of research at UNamur. 

Terminator to talk about AI? Wall-E to talk about technological dependence? The Truman Show to discuss social media? In a new book, two professors from UNamur, Anthony Simonofski (digital transformation—EMCP Faculty—NaDI Institute) and Benoît Vanderose (software engineering—Faculty of Computer Science—NaDI Institute), take readers on a journey at the crossroads of digital technology and cinematic imagination. 

Our research activities are divided into 3 main areas.

Since 2017 and as part of the work of the Pacte pour un Enseignement d'excellence, eight scientific "Consortiums" of the tronc commun have been formed. They bring together a total of some fifty members from the various universities, colleges and higher arts schools in the Wallonia-Brussels Federation. Their mission is to establish, on the e-classe digital platform, a repertoire of didactic tools and pedagogical devices that the educational community will be able to consult, select and procure easily, according to its needs and those of its students. These resources are available on the e-classe digital platform.

A Hackathon is a short event where teams reflect on a particular theme. They attempt to find solutions by developing innovative projects. At the end of the event, a jury determines which projects have caught their attention and thus won the competition. A CSLabs initiative The Computer Science Labs is a junior enterprise that grew out of the Faculty of Computer Science at the University of Namur. In practical terms, its actions revolve around carrying out projects, training members on IT-related topics and organizing events. Read more