Hugues Talbot

Hugues Talbot is a researcher affiliated with the Digital Vision Center, specializing in computer vision, machine learning, and optimization. His work spans medical imaging, digital pathology, and mathematical morphology, with recent publications focusing on deep learning applications for tumor segmentation, image restoration, and medical diagnostics.

• Research Interests: Computer Vision (e.g., image segmentation, mathematical morphology), Machine Learning (e.g., generative models, multiple instance learning), Medical Imaging (e.g., CT/MRI analysis, tumor detection), and Optimization (e.g., maximal flows, convex modeling).
• Key Projects: Development of algorithms for automated lymphoma prognosis using PET images, deep learning models for breast cancer diagnosis from cytology slides, and physics-informed neural networks for myocardial perfusion simulation.
• Technical Contributions: Hybrid heart segmentation algorithms, noise density analysis for image splicing detection, and discrete calculus methods for inverse problems in imaging.

His work often integrates theoretical mathematical models (e.g., Ambrosio-Tortorelli functional) with practical applications in clinical settings, such as sacroiliitis detection, bone density estimation, and zebrafish brain volumetric analysis.