Image restoration: from model to learning based techniques [1 ECTS MATH05]

This course introduces image reconstruction through theoretical and practical sessions. Participants will study image formation and degradation models, classical and learning-based reconstruction methods, and hybrid approaches such as Plug-and-Play, algorithm unrolling, and deep priors. Lectures and laboratory sessions provide a comprehensive overview of traditional and modern techniques for image reconstruction.

• 9 April 10:30 - 12:30, Room H: Introduction, image formation model, image degradations models and causes, problem formulation. Introduction to image reconstruction methods. Image formation, classical image reconstruction methods.
• 13 April 12:30- 14:30, Room H: Model-based and learning based approaches. Image reconstruction based on machine learning.
• 16 April 10:30 - 12:30, Room H: Hybrid approaches (Plug&Play, Algorithm unrolling, Deep priors...). Hybrid image reconstruction methods.

Further information: giacomo.albi@univr.it