area: Rouen, France
Team: Quantif, LITIS laboratory, University of Rouen, https://www.litislab.fr/equipe/index/quantif
Duration: 12 months
Start date: from beginning of 2020
In the field of radiological oncology, particularly in the case of individualized radiotherapy, the segmentation of organs and tumor from medical images are primordial. Artificial intelligence (AI) is revolutionizing many fields thanks to its new learning capabilities on large databases. However, for medical imaging, the requirements of learning from large databases are usually unachievable. Even if we could have a large database, the amount of data would be too large to be annotated, because the annotation of medical data is long and tedious and can only be done by physicians. Therefore, the challenge is to increase learning capacity with a small amount of annotated data.
Deep learning is currently a very active area of research, particularly in medical imaging. The objective of this work is to study and develop semi-supervised deep learning methods by combining a priori knowledge in case we have little annotated data. The application will focus on the semantic segmentation of medical images to improve radiotherapy.
Due to its multidisciplinary orientation combining medicine and information sciences, the work will address the issues raised both for health and also for the development of new technologies. It will be carried out in close collaboration with the hospital center (CLCC) Normandy – Rouen (Henri Becquerel Center) which will provide images and medical expertise.
Skills and Profile
- PhD in computer science, data science, image processing (medical image processing), computer vision or applied mathematics.
- Excellent programming skills.
Su Ruan <firstname.lastname@example.org>, Caroline Petitjean < email@example.com >