Organizing Committee
Comité d’organisation
Luca Calatroni (Università di Genova)
Emmanuel Soubies (CNRS, IRIT Toulouse)
Pierre Weiss (CNRS, IRIT Toulouse)
The workshop will encompass a thorough exploration of topics at the intersection of physical modeling and computational imaging. Thematic sessions will range from the foundations of (blind) inverse problems with both analytical and geometrical viewpoints to the use of advanced optimization techniques tailored to incorporate physical constraints into image reconstruction algorithms. Scientific presentations and hands-on sessions describing how efficient artificial intelligence techniques and software can be used in this area to enhance imaging performance will also be organized. A wide range of applications of these methodologies will be explored, with a particular focus on optical microscopy where precise imaging is crucial for scientific discoveries at the nanoscale.
L’atelier comprendra une exploration approfondie des sujets à l’intersection de la modélisation physique et de l’imagerie informatique. Les sessions thématiques iront des fondements des problèmes inverses (aveugles) avec des points de vue analytiques et géométriques à l’utilisation de techniques d’optimisation avancées conçues pour incorporer des contraintes physiques dans les algorithmes de reconstruction d’images. Des présentations scientifiques et des sessions pratiques décrivant comment des techniques et des logiciels d’intelligence artificielle efficaces peuvent être utilisés dans ce domaine pour améliorer les performances de l’imagerie seront également organisées. Un large éventail d’applications de ces méthodologies sera exploré, avec un accent particulier sur la microscopie optique où l’imagerie précise est cruciale pour les découvertes scientifiques à l’échelle nanométrique.
SPEAKERS
Andres Almansa (Université Paris Cité) Accelerating Posterior Sampling with Generative Priors for Blind Inverse
Hilton Barbosa de Aguiar (ENS Paris) Single-pixel methods for linear and nonlinear microscopy
Laure Blanc-Féraud (Université Côte d’Azur) Off-the-Grid Curve Reconstruction in blurred Images by optimization of functional
Nathan Buskulic (University of Genova) Learning the optimal Tikhonov regularization for blind inverse problem, what should you expect ?
Caroline Chaux (CNRS, I2M) Learning Weighted Least Squares Through Unrolling for Poisson Image Deconvolution
Paola Causin (University of Milano) Manifold Learning Approaches via Riemannian Geometry: Application to Inverse Problems in Biomedical Imaging
Lisa Cuneo (Italian Institute of Technology) & Luca Calatroni (University of Genova) Reconstruction Approaches in Image Scanning Microscopy: Regularization and Optimization
Christian Daniele (University of Genoa) Deep Equilibrium Models for Poisson Inverse Problems via Mirror Descent
Julie Delon (Université Paris Cité) From distributions to flows: an introduction to Flow Matching
Valentin Debarnot (CREATIS, INSA Lyon) Supervised learning in tomography with partially unknown forward operator and no ground- truth
Charles Dossal (INSA Toulouse) Inertia as a preconditioner
Fabian Erdel (CNRS, CBI Toulouse) Biomolecular condensates: Structure, dynamics and mechanics
Paul Escande (Institut de Mathématiques de Toulouse) On the numerics of photoacoustic tomography
Jalal Fadili (CNRS, ENSICAEN) Inertial Algorithms Meet NN-Based Methods for Inverse Problems
Rémi Gribonval (ENS de Lyon) Training dynamics of ReLU Networks: a Path-lifting Perspective
Johannes Hertrich (Université Paris Dauphine) Importance Corrected Neural JKO Sampling
Samuel Hurault (ENS – PSL) From Denoising to Diffusion : A Fine-Grained Error Analysis
Laurent Jacques (UCLouvain) Herglotz-NET: Implicit Neural Representation of Spherical Data with Harmonic Positional Encoding
Oihan Joyot (IRIT) Learning the Dynamic Law of Cellular Condensates
Ulugbek Kamilov (Washington University) Computational Imaging: Restoration Deep Networks as Implicit Priors
Felix Krahmer (Technical University of Munich) Solving Inverse Problems with Deep Linear Neural Networks: Global Convergence Guar- antees for Gradient Descent with Weight Decay
Hippolyte Labarrière (University of Genoa) Reparameterization and Its Role in Optimization Dynamics
Tobias Liaudat (CEA-Saclay) Point spread function wavefront recovery: phase retrieval with automatic differentiation
Amol Mahurkar ( ENS Paris) Generalized Energy Sensing for Imaging
Thomas Mangeat (Centre de Biologie Intégrative Toulouse) RIM and PRIM for subcellular dynamics on tissue or medium content screening.
Ségolène Martin (TU Berlin) & Anne Gagneux (ENS de Lyon) TBA
Mathurin Massias (INRIA) On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
Giacomo Meanti (INRIA) Unsupervised Imaging Inverse Problems with Diffusion Distribution Matching
Hadrien Montanelli (INRIA) Neural networks for inverse scattering
Thomas Moreau (INRIA) Filling the gaps: a story of priors and conditional probabilities
Nathanaël Munier (Université Toulouse III-Paul Sabatier) Jackpot: Approximating Uncertainty Domains with Adversarial Manifolds
Minh Hai Nguyen (Université Toulouse III – Paul Sabatier) How Diffusion Prior Landscapes Shape the Posterior in Blind Deconvolution
Nicolas Papadakis (Institut de Mathématiques de Bordeaux) Posterior Sampling with the Proximal Stochastic Gradient Langevin Algorithm
Romain Petit (CNRS, ENS) On the non-convexity issue in electrical impedance tomography
Audrey Repetti (Heriot-Watt University) TBA
Florian Sarron (Université Toulouse III – Paul Sabatier) Learning to identify PSFs in fluorescence microscopy
Anne Sentenac (CNRS, Institut Fresnel) Super-resolved fluorescence microscopy using random illuminations (RIM)
Emmanuel Soubies (CNRS, IRIT) Introduction to super-resolution fluorescent microscopy
Matthieu Terris (INRIA) Reconstruct Anything Model: generalizing restoration models beyond a single task
Pauline Trouvé-Peloux (ONERA) End-to-end design of imaging systems and neural networks – application to the design of a privacy preserving camera
Hervé Turlier (CNRS, Collège de France) Bridging Fluorescence Microscopy and Biophysical Tissue Models
Samuel Vaiter (laboratoire jean-alexandre dieudonné) Remarks about bilevel optimization
Pierre Weiss (IRIT) Analytical solutions for CNN inverse problem solvers
SPONSORS
