‹ Back to the directory

Jesus Angulo Lopez

Jesus Angulo Lopez

Research Director

Center · CMA

Biography

Jesus Angulo Lopez is a researcher specializing in image processing and deep learning, with particular expertise in mathematical morphology and equivariant neural networks. His work focuses on integrating morphological operators into deep learning architectures, particularly to address issues related to symmetries (rotations, translations, scale changes) and model robustness. He also explores theoretical approaches to generalizing classical morphological operators—such as erosion and dilation—within advanced algebraic frameworks (non-Abelian groups, (max,+) algebra) and their differentiable approximations for use in neural networks. His research includes applications in medical imaging, radar anomaly detection, and texture analysis, while contributing to the mathematical foundations for equivariant and interpretable architectures.

Publication(s)

Teaching

Image Analysis: From Theory to Practice

Lecturer

The course includes a theoretical presentation on image filtering and segmentation tools in the morning and tutorials in the afternoon. Binary, grayscale, and color images, as well as 2- and 3-dimensional images, are covered. Examples drawn from a wide variety of real-world applications illustrate the capabilities of the tools presented.

Color, Arts, Industry

Course Director

Perceptual and Physical Approaches to Light and Color Practical Exercises on Color Color Reproduction in the Digital Image Processing ChainDigital Image Processing Mathematics of digital color image processing The Color of Minerals and Color Minerals Virtual Reality From Digital Photography to the 1842 Cyanotype Discovering Color Minerals at the Museum of Mineralogy Color in an Industrial Context Group Project Supervised sessions are set aside for work on the group project.

ATHENS - MP08 - Physics and Mechanics of Random Media

Guest Lecturer

Based on a review of advanced experimental techniques for describing microstructure, and on typical results involving fluctuations in the phenomena of plasticity, damage, fracture, and flow in porous media, the basic tools of applied probability and random processes are reviewed. Probabilistic tools for describing random media and models, as well as their simulation, are introduced. The physics and mechanics of random media are first presented from the perspective of approximate solutions to partial differential equations with random coefficients. For example, problems in linear electrostatics in random media are studied using a perturbation expansion of random electric and displacement fields, while the limits of the effective permittivity and elastic moduli are derived from variational principles. This homogenization approach, which can be applied to other physical properties such as permeability or thermal conductivity, is illustrated by third-order bounds. The use of numerical techniques (such as the finite element method) to estimate the homogenized properties of random media based on Monte Carlo simulations is introduced. The limits and numerical techniques are then extended to nonlinear behaviors, such as the plasticity of polycrystals. Given the importance of reliability issues in a wide range of engineering applications, several statistical failure models (brittle, ductile, fatigue) are developed using a probabilistic approach. Course Structure: One week. Lectures (80%) and hands-on computer training (20%) More information at: http://cmm.ensmp.fr/Enseignement/es-physrandmedia.html

PhD supervision

  • 2022 Ensemble model construction automation and hyperparameter optimization MAILLET Vassili
  • 2021 Contributions to unsupervised visual anomaly detection CASAGRANDE BERTOLDO Joao Paulo
  • 2021 Anomaly detection in satellite trajectories using artificial intelligence for space surveillance radar BAUDIER Stéfan
  • 2020 Contributions to Equivariance under Roto-Translations for Deep Learning in Image Processing PENAUD--POLGE Valentin
  • 2019 One-class classification for low-resolution low-supervised Doppler pulse radar target discrimination BAUW Martin
  • 2019 Deep Learning Equivariant Based on Scale Spaces and Moving Frames SANGALLI Mateus
  • 2019 Characterization of the physicochemical properties of steel surfaces by hyperspectral imaging. ZENATI Tarek
  • 2016 Artificial learning for image segmentation CHANG Kaiwen