Home »

Proposition de stage - 2024

Statistical-shape models for the assistance of lower-limb muscles

Niveau : Engineer or MSc. INTERNSHIP POSITION

Context: the FULGUR Project

This internship position is part of the FULGUR project led by Gaël GUILHEM (Laboratory Sport Expertise and Performance, INSEP), supported by the national research agency program “Programme Prioritaire de Recherche Très Haute Performance Sportive” (Grant for research applied to very high performance in sports), within the framework of the Olympic/Paralympic Games that will take place in Paris in 2024. FULGUR is a multi-disciplinary program on high performance in explosive sports that include 9 research centers, 3 sports federations (Athletics, ice sports of bobsleigh and rugby) and 2 companies.

One objective of this project is to determine the individual musculoskeletal profile of athletes in order to propose individualized training programs (WP 2). WP2, led by Antoine NORDEZ (Full Professor at the Université de Nantes, IUF, France) and Giuseppe RABITA (Researcher at the INSEP, Paris), aims to investigate the relationships between muscle coordination and
determinants of sprint propulsion performance. A previous internship and engineering project led to the development of a tool to automatically segment 18 muscles of the lower-limb with Deep-learning methods [3]. The segmentations allowed to estimate the volume of the muscles mainly involved in sprinting performance and helped for training individualisation
and injury prevention.

Internship GOALs

In continuation of the automatic segmentation project, the intern’s role will largely involve ensuring the smooth progress of this project:
● Applying inference to the entire FULGUR database, thus enabling a statistical study of the impact of certain parameters on muscle volume (age, gender, height, weight, body mass index) and comparing it with the literature [4].
● Conducting an uncertainty study on the predictions based on [1],[2].
● Implementing a shape statistical analysis on the results obtained through ShapeWorks exploration.


1. The student should be enrolled in a Master (M1 or M2) or in the last year of anEngineering program, in one of the following (or related) domains: Machine/Deep Learning, Signal and Image Processing, Applied Mathematics, or Computer vision.
2. Prior knowledge in Python is a must. A minimum level of experience in deep learning libraries such as Tensorflow, Keras or Pytorch is required.
3. Knowledge in image analysis, specially medical would be a plus.

Personnes à contacter
Fiche descriptive :

Partager ce contenu
Copyright : LS2N 2017 - Mentions Légales -