[Trainee] – Deep Learning for Infant Motion Classification
Inria
Montbonnot-Saint-Martin, 38330
[Trainee] – Deep Learning for Infant Motion Classification
Le descriptif de l’offre ci-dessous est en Anglais
Niveau de diplôme exigé : Bac + 3 ou équivalent
Fonction : Stagiaire de la recherche
A propos du centre ou de la direction fonctionnelle
Grenoble Rhône-Alpes Research Center groups together a few less than 800 people in 39 research teams and 8 research support departments.
Staff is localized on 5 campuses in Grenoble and Lyon, in close collaboration with labs, research and higher education institutions in Grenoble and Lyon, but also with the economic players in these areas.
Present in the fields of software, high-performance computing, Internet of things, image and data, but also simulation in oceanography and biology, it participates at the best level of international scientific achievements and collaborations in both Europe and the rest of the world.
Contexte et atouts du poste
Context:
Medical motion analysis is a valuable clinical tool for evaluating an individual’s state of health. One of the most common motor disorders in childhood is cerebral palsy (CP), which can be detected by means of infant motion analysis [1]. Trained experts can identify infants at high risk of developing CP solely based on the assessment of general movements (GMA) [2], i.e., they evaluate the quality of general movements (GMs) from videos of spontaneously moving infants. GMA provides excellent reliability, but becoming a skilled GMA expert requires time and regular practice [1]. An automation of GMA can support clinicians in identifying high-risk infants as early as possible in order not to miss an opportunity for early therapy.
Mission confiée
Objectives
The goal of the internship is to be able to predict the GMA rating of an infant from a sequence of 3D poses. The method should, in addition to give a prediction, also state its associated confidence. Additionnaly, the exploration of the activation layers should provide an identification of which specific motions are relevant for the classification.
First, a study of how existing architectures perform on the classification task will be conducted, by focusing on the transformers architectures. Then, the intern will explore new variants / architectures to improve the state of the art results.
Data corpus
An internal dataset is available containing the motion sequences of the infants and their medical GMA obtained with [3,4].
References:
[1] Hadders-Algra, M.: General movements: a window for early identification of chil- dren at high risk for developmental disorders. The Journal of Pediatrics 145, S12 – S18 (2004)
[2] Prechtl, H.: Qualitative changes of spontaneous movements in fetus and preterm infant are a marker of neurological dysfunction. Early Human Development 23(3), 151 – 158 (1990)
[3] Nikolas Hesse, Sergi Pujades, Javier Romero, Michael Black, Christoph Bodensteiner, Michael Arens, Ulrich Hofmann, Uta Tacke, Mijna Hadders-Algra, Raphael Weinberger, Wolfgang Müller-Felber, A. Sebastian Schroeder. Learning an Infant Body Model from RGB-D Data for Accurate Full Body Motion Analysis. In A. Frangi, J. Schnabel, C. Davatzikos, C. Alberola-López, & G. Fichtinger (Eds.), Medical Image Computing and Computer Assisted Intervention, MICCAI, pp.792-800, 2018
[4] Nikolas Hesse, Sergi Pujades, Michael Black, Michael Arens, Ulrich Hofmann, Sebastian Schroeder. Learning and Tracking the 3D Body Shape of Freely Moving Infants from RGB-D sequences. IEEE Transactions on Pattern Analysis and Machine Intelligence, Institute of Electrical and Electronics Engineers, 2019, pp.12. ⟨10.1109/TPAMI.2019.2917908⟩
Principales activités
Observe, think, act
Compétences
Candidate Profile:
– Master student – preferably in Computer Science or Applied Mathematics.
– Creative and highly motivated
– Solid programming skills
– Fluent English or French spoken.
– Prior courses or knowledge in the areas of temporal series, machine learning and deep learning are a plus
Avantages
- Subsidized meals
- Partial reimbursement of public transport costs
- Professional equipment available (videoconferencing, loan of computer equipment, etc.)
- Social, cultural and sports events and activities
Rémunération
Gratification minimum légale