Day 2

Tuesday, July 2

• 2 lectures on radiomics and artificial intelligence (AI)
• 2 dedicated practical sessions dedicated to AI
• 1 spot on nuclear medicine
• Gala dinner in Nantes city center

• Introduction to Artificial Intelligence (AI)
  Diana Mateus, LS2N
  Professor at Centrale Nantes, Research Scientist in Medical Image Analysis

The presentation will start by an introduction to the basics of machine learning. Then, the focus will shift towards deep learning methods for the classification and segmentation of medical images, highlighting the specific challenges of nuclear imaging.

• Radiomics
  Thomas Carlier, CRCI²NA
  Medical Physicist at the Nuclear Medicine Department, University Hospital of Nantes

Radiomics is an advanced field of medical imaging that involves the extraction of a large number of quantitative features from medical images using data characterization algorithms. In nuclear medicine, radiomics can play a crucial role by analyzing images obtained from positron emission tomography (PET) and, eventually single-photon emission computed tomography (SPECT).

In this talk, we will start from the early beginnings of radiomics to the most advanced techniques in nuclear medicine, including engineered features and deep features, with a small emphasis on the concept of harmonization. Some recent convincing examples will be given.

• Image processing and deep learning

Session 1 : Computation of radiomics features
Session 2 : Image segmentation with deep learning

These sessions will demonstrate through interactive programming tools how to :

• extract radiomics features from images
• show the internal mechanism of neural networks
• collect expert knowledge in the form of image annotations useful to train machine learning tools

• Artificial intelligence in nuclear medicine

• Back to ISI NucMed program 2024