PhD position- Oct 2018 - Machine learning for cancer screening based on multimodality imaging
Recrutement en cours/passé: 
Recrutement passé

Please see the attached file for a detailed description of the project.

Scientific fields : machine learning, deep learning, multimodality imaging, clinical decision systems

Background :CREATIS has developed strong skills in developing clinical decision systems (CAD) for cancer and brain imaging based on the most advanced machine learning techniques. Such systems are designed to assist clinicians in their diagnosis by highlighting abnormal regions in an image. One active project of the ‘Images and Models’ team concerns the prototyping of a computer-aided diagnosis system for prostate cancer screening based on multiparametric magnetic resonance imaging (MRI). Despite an important improvement brought by such systems for the problem of cancer mapping, they still suffer from limitations that restrict them from being used at a larger scale.

Project aim : The purpose of the PhD project is to go one step further and address two challenges:

  • to predict not only a presence/absence of cancer but also the degree of its aggressiveness. The main challenge is that the different classes, corresponding here to the different levels of lesion aggressiveness, are highly correlated and interdependent which challenges standard multi-class classification algorithms.
  • to develop a system whose performance generalize well with data potentially coming from different populations, as can be encountered with imaging data pooled over different clinical centers

From the methodological point of view, we plan to explore new machine learning algorithms that tackle the problem induced by the presence of highly correlated and interdependent outcomes in multi-class classification as well as heterogeneous data. One research axis that will be investigated is to explore the potential of deep learning to address both questions. Our objective will be to investigate novel deep architectures that will efficiently fit our needs, particularly focusing on semi-supervised networks allowing to operate on partially labeled data, which is a major characteristic of medical data.

Skills : The candidate is expected to have strong knowledge either in machine learning or image processing and a good experience in both fields. Some prior experience with medical image processing would be appreciated but is not required. Good programming skills are also required. The available code is written in Matlab and Python but other languages can be used. We are looking for an enthusiastic and autonomous student with strong motivation and interest in multidisciplinary research (image processing and machine learning in a medical context).

Salary : Around 1700 euros net salary (+ teaching)

Applications : Interested applicants are required to send a cover letter, CV and any other relevant documents (reference letter, recent transcripts of marks,...)  before June 22 to: