Close

‹ Back to the directory

Etienne Decenciere

Etienne Decenciere

Research Director

Center · STIM

Discipline(s)
Signal, Image, Automatic Control, Robotics and Industrial Engineering
Topic(s)
Biodiversity, Medical Image Analysis

Keywords

Team

STIM

Biography

Etienne Decencière is a researcher specializing in image processing and analysis, with significant expertise in applying advanced computer vision and deep learning methods to a variety of fields, such as composite materials, medical imaging, and biology. His work focuses on developing automated tools for the quantitative analysis of complex structures, particularly through approaches based on convolutional neural networks (CNNs), generative models, and mathematical morphology techniques. His research covers topics such as 3D image segmentation, anomaly detection in medical data (retinal, histopathological), and the optimization of microstructures for applications in materials science. A significant portion of his work aims to improve the robustness, interpretability, and efficiency of models, while reducing the need for annotated data—a key challenge for clinical and industrial applications. His methods, often validated on public or experimental datasets, demonstrate a pragmatic approach that combines mathematical rigor with adaptation to the real-world constraints of the application domains.

Publication(s)

Projects

  • 2025-2028 Classification de plancton d'eau douce Lead Investigator Développement de méthodes de classification du plancton d'eau douce. Financement TTI.5. Collaboration avec le CEREEP Ecotron Idf. Co-directeur de thèse: Jean-François Le Galliard
  • 2025-2028 Analyse de vidéos acoustiques Participant Analyse de vidéos acoustiques pour limiter l'impact de la production hydroélectrique sur la migrations de poissons. Thèse de Fabian Roldan Figueredo.

Teaching

Deep Learning for Image Analysis

Course Director

The course will consist of lectures and hands-on sessions. The lectures will cover: - Introduction to machine learning. - Artificial neural networks, backpropagation algorithm - Convolutional neural networks - Successful architectures - Analysis of neural network functions - Image classification and segmentation - Autoencoders and generative networks - Current trends and research prospects During the practical sessions, students will implement the concepts learned in class using Python. They will tackle practical problems related to deep learning: architecture design, optimization strategies, hyperparameter selection, and analysis of results.

Data, Images, Physical Models, and Machine Learning (Research Quarter)

Course Director

Data analysis is playing an increasingly important role in our society, both professionally and personally. This data is often complex: text, images, videos, genomes, point clouds, and graphs, for example. It serves as the raw material for the digital industries. Automatically extracting useful information from these massive datasets is a major challenge. The goal of this research term is to provide engineering students with their first experience in research on the automatic analysis of complex data. Images and other structured data (graphs, trees, sequences, etc.) will be the focus of this work. The scientific disciplines involved will include, in particular, machine learning, image analysis, robotics, physics, and bioinformatics. Depending on the projects, students will have the opportunity to use kernel methods, deep learning, or mathematical morphology, among other techniques. The fields of application will be very diverse, ranging from autonomous driving to healthcare, including non-destructive testing and materials characterization. Some research projects will be conducted in collaboration with industry partners.

PhD supervision

  • 2025 Development of methods for analyzing acoustic videos to identify and quantify fish in natural environments around energy infrastructures. ROLDAN FIGUEREDO Xabier Fabian
  • 2025 Generative artificial intelligence models for metallic alloy microstructures COURTOIS Martin
  • 2025 Automatic classification of lacustrine plankton for characterizing ecosystem functioning DÉCHAUMET Léo
  • 2023 Dynamic and morphological characterization of the electrochemical environment of lithium-ion batteries during charge cycles BOTTENMULLER Antoine
  • 2023 Machine learning-based estimation of molecular biomarkers from H&E images: characterization of performance, limitations, and downstream applications BALEZO Guillaume
  • 2022 Advances in 3D vision and deep learning for shape analysis and synthesis: application to interspecies biomechanical modeling of the knee joint BASTICO Matteo
  • 2022 Deep learning approaches for the analysis and optimization of fiberglass-reinforced polymer matrix composites BASSO DELLA MEA Guilherme
  • 2021 Exploitation of fundus images for retinal biometrics using deep learning LANGROGNET Thomas
  • 2021 Contributions to unsupervised visual anomaly detection CASAGRANDE BERTOLDO Joao Paulo
  • 2021 Deep learning methods for sparse-view X-ray tomography VO Romain
  • 2019 Computational pathology representation learning: application to predicting cancer molecular features LAZARD Tristan
  • 2015 Image segmentation problems and contribution to mathematical morphology ALAIS Robin