Context: The project deals with diesase which is a major international health concern and that is manifested by visible clinical signs. The goal is to study a “pre-diagnosis” system able to analyze “standard” 2D photographs taken with consumer digital cameras (ex: Smartphone) in order to predict whether the patient should see a doctor. Given a training collection of images annotated according to the presence or the absence of a pathology and its level of development if present, the system should be trained to be able make a pre-diagnostic on unseen images. Visible clues are also annotated and should be detected too in order to justify the system's decision. This internship will take plac in the MRIM team but it will be in the context of a cooperation with a start-up that will provide the training and test images, as well as all the expertise related to the targeted pathology.
Objective: the objective of the project is to produce a system able to do the pre-diagnosis task but it is also to provide explanations regarding how the conclusion was reached. For this, the system should be able to identify the types of elements or of attributes that it uses for making the decisions and how these are used for that, ideally in terms of visual clues and of logical rules on them. The scientific aspects will be mostly related to the explainability part while classical deep learning-based methods will be used for the decision part.