Publications

Image Processing

S3VD SELF-SUPERVISED SPATIAL VIDEO DOWNSAMPLING LOSS: A METHOD FOR TRAINING VIDEO FPN DENOISING NETWORKS

Fixed pattern noise (FPN) is a temporally constant noise present on videos due to the non-uniformities of the sensors that may exhibit spatial correlation, typically across columns and/or rows. Acquiring real clean/noisy data is particularly…

Hortensia Barral, Pablo Arias, Axel Davy

The IEEE International Conference on Image Processing (ICIP2025) - 2025

Dermatology

Chiasmal Decussation in Oculo-Cutaneous Albinism Type 8

PURPOSE. Albinism is a genetic disorder characterized by a defect in melanin biosynthesis. Ophthalmological and dermatological impairments vary according to the patient genotype and are highly heterogenous. Recently, variants in the DCT gene were…

Maxence Rateaux, Smaïl Hadj-Rabia, Rémi Barrois, Olivia Zambrowski, Vincent Michaud, Ester Moreno-Artero, Dominique Bremond-Gignac, Sophie Javerzat, Matthieu P Robert

Investigative Ophthalmology & Visual Science - 2025

Analysis of PDEs

Existence, instability and energy partition for linear or nonlinear evolution equations issued from fluid mechanics

This thesis contains three themes: microlocal partition of energy, the virial theorem, and the well-posedness of jet flows. The first theme explores how the energy of solutions to dispersive equations is equally distributed inside and outside the…

Haocheng Yang

Differential Geometry

Barycentric subspace analysis of network-valued data

Certain data are naturally modeled by networks or weighted graphs, be they arterial networks or mobility networks. When there is no canonical labeling of the nodes across the dataset, we talk about unlabeled networks. In this paper, we focus on the…

Elodie Maignant, Xavier Pennec, Alain Trouvé, Anna Calissano

Image Processing

Pair-dependent representations for Same-View Vehicle recognition

Deep metric learning approaches for object instance recognition struggle with generalization. To infer if two images are of the same instance, a representation is computed for each image, independently of the other. This is insufficient for reliably…

Anis Yassine Ben Mabrouk, Antoine Tadros, Axel Davy, Rafael Grompone von Gioi, Gabriele Facciolo

Analysis of PDEs

Numerical study of the Breaking of Short-Crested Waves : Breaking Onset, Energy Dissipation and Superharmonic Instability

Wave breaking plays a crucial role in energy dissipation and turbulent mixing at the ocean surface. Accurate predictions require robust breaking criteria and dissipation models. While most studies focus on two-dimensional systems, this thesis…

Ayoub Mansar

Numerical Analysis

Numerical modeling accidental release of cryogenic liquid : application to liquid hydrogen (LH2) risk assessment

In the face of energy shortages and increasing pollution, many countries are developing policies to replace fossil fuels with clean alternative energy sources. Storage and transportation are crucial in the hydrogen sector. Compression and…

Yanwei Liang

Artificial Intelligence

OneBatchPAM: A Fast and Frugal K-Medoids Algorithm

This paper proposes a novel k-medoids approximation algorithm to handle large-scale datasets with reasonable computational time and memory complexity. We develop a local-search algorithm that iteratively improves the medoid selection based on the…

Antoine de Mathelin, Nicolas Enrique Cecchi, François Deheeger, Mathilde Mougeot, Nicolas Vayatis

Proceedings of the 39th annual AAAI Conference on Artificial Intelligence - 2025

Probability

Conformal Approach to Gaussian Process Surrogate Evaluation with Coverage Guarantees

Gaussian processes (GPs) are a Bayesian machine learning (ML) approach widely used to construct surrogate models for the uncertainty quantification (UQ) of computer simulation codes in industrial applications. It provides both a mean predictor and…

Edgar Jaber, Vincent Blot, Nicolas Brunel, Vincent Chabridon, Emmauel Remy, Bertrand Iooss, Didier Lucor, Mathilde Mougeot, Alessandro Leite

Journal of Machine Learning for Modeling and Computing - 2025

Environmental Sciences

Progress and trends in liquid hydrogen release

Liquid hydrogen (LH2) is a promising method for large-scale, long-distance storage and transportation of green energy. However, LH2 leaks pose significant safety risks, such as fire and explosion, due to the rapid formation of gaseous hydrogen…

Yanwei Liang, Ming He, Yongfeng Qu, Nan Peng, Jean-Michel Ghidaglia, Liqiang Liu

International Journal of Hydrogen Energy - 2025