Doctoral lecture in Electrical and Computer Engineering - Matthieu Muller

Doctoral candidate: Matthieu Muller

Title of thesis: Hybrid Deep Learning Methods for Image Reconstruction: Application to Demosaicking

Opponents:
Dr. Paolo Gamba, Professor at the University of Pavia, Italy.
Dr. Jon Yngve Hardeberg, Professor at the Norwegian University of Science and Technology, Norway.

Advisor: Dr. Magnús Örn Úlfarsson, Professor at the Faculty of Electrical and Computer Engineering, University of Iceland.

Other members of the doctoral committee:
Dr. Mauro Dalla Mura, Professor at the University Grenoble-Alpes, France.
Dr. Daniele Picone, Postdoc at the University Grenoble-Alpes,France.

Head of Ceremony: Dr. Lotta María Ellingsen,  Professor and Head of the Faculty of Electrical and Computer Engineering, University of  Iceland. 

This is a joint doctoral degree with the University Grenoble-Alpes, France. The doctoral defense took place in France on December 10th.

Abstract

Image processing is essential for extracting meaningful information from sensor-acquired visual data, with applications ranging from medical imaging to remote sensing. Raw images often suffer from noise, blur, limited resolution, and CFA artifacts due to sensor and environmental constraints, making restoration a critical step. Traditional model-based approaches, grounded in physical and mathematical principles, provide interpretability but often fail to capture the complexity of real-world degradations. Deep learning has recently revolutionized image restoration by learning powerful mappings from degraded to clean images, though it requires large labeled datasets, significant computational resources, and offers limited interpretability. Hybrid approaches that combine model-based techniques with the flexibility of neural networks aim to overcome these limitations by improving robustness, reducing data demands, and providing more explainable solutions. This thesis explores such hybrid methods, with a particular focus on the universal demosaicking problem, that is reconstructing full-color images from gray-scale CFA inputs.

About the doctoral candidate

Matthieu Muller received the M.Sc. degree in Applied and Industrial Mathematics from the University Grenoble-Alpes, Grenoble, France in 2022, along with an Engineering degree from the Grenoble INP - Ensimag, Grenoble, France engineering school the same year. His research interests include signal and image processing, deep learning, hybrid methods for image processing, self-supervised learning, and remote sensing.