Advancing Brain Tumor Segmentation through Deep Learning based Approaches

We propose a semi-automated brain lesions segmentation method utilizing deep learning techniques and point prompting such as foundation models (SAM).
This developed approach aims to assist physicians in the follow-up of multiple sclerosis disease at different stages, enhancing the likelihood of successful treatment for patients. The key contribution of this research lies in the advancement and implementation of a brain lesion segmentation method, ultimately improving the quality of life for individuals who suffer from multiple sclerosis.
Our project employs image processing techniques and Deep Neural Networks (DNN) to achieve this goal. Should the developed project prove as valuable as we hope, it has the potential to better assess the pathology evolution of patients, signifying its substantial impact on patient well-being which is beneficial for both participating institutions. Furthermore, we are proposing cutting-edge research, which has significant benefits for both participating institutions, enhancing their academic reputation and fostering innovation that can lead to practical applications.

Faculty Supervisor:

Thomas Fevens

Student:

Partner:

Institut National des Sciences Appliquées de Lyon

Discipline:

Computer science

Sector:

Artificial Intelligence; Health and Related Sciences & Technology; Information and Communications Technology

University:

Concordia University

Program: