Academic work


Combining medical imaging and
motion capture for the dynamic analysis
of osteoarticular structures

Privat-docent Thesis

This thesis presents our interdisciplinary approach and the methodology implemented in our research, as well as the objectives and challenges to be addressed in order to combine medical imaging and motion capture to study and dynamically analyze personalized osteoarticular structures. The second part of this thesis presents a selection of our most significant work in the field that illustrates concretely how this methodology was implemented to answer clinical questions in the framework of diagnosis, treatment and surgical planning of orthopedic disorders. This work is organized around the 3D modeling and simulation of three particular joints: 1) the hip, 2) the shoulder complex and 3) the knee. This thesis includes scientific and clinical studies focusing on the biomechanical modeling of human joints and on the exploration of several motion-related disorders, such as femoroacetabular impingement in the prosthetic hip, knee ligaments biomechanics, glenohumeral instability and impingement, rotator cuff hyperelongation, hip and shoulder implants design, etc.


Extreme hip movements based on
optical motion capture

PhD Thesis

Professor: Nadia Magnenat-Thalmann
Jury: J. Rolim, F.-E. Wolter, M. de Zee, E. Stindel

Many causes can be at the origin of hip osteoarthritis, but the exact pathogenesis for idiopathic osteoarthritis has not yet been clearly delineated. Indeed, changes in the movement and alignment of the hip could be other potential causes of early osteoarthritis. In particular, athletes seem to present a higher risk of developing such disease due to repetitive and extreme movements performed during their daily activities. In this thesis, a clinical study with professional ballet dancers is being conducted. The goal of this study is to verify if repetitive extreme movements can be a factor of hip joint degeneration through joint subluxation and excessive cartilage deformations. To achieve these goals, we developed a motion protocol using optical motion capture system to estimate the kinematics of patient-specific hip joint 3D models. We also developed an effective and robust correction method to minimize soft tissue artifacts, a critical issue in human motion analysis.

Virtual Mirror: A real-time motion
capture application for

Master Thesis

In collaboration with Clémentine Lo
Professor: Nadia Magnenat-Thalmann
Assistant: Etienne Lyard

The goal of this project is to create a 3D application which will give the illusion of a “virtual mirror” by using real-time motion capture. As MIRALab is using OpenSceneGraph (OSG), a graphic library, to develop its real-time cloth animation application, the same toolkit is employed for this project. The subject’s movements are tracked using a Vicon optical motion capture system and are applied to a virtual avatar in real-time. The avatar is included in a 3D environment and the scene is projected on a big screen in front of the subject. The camera projection is calculated so as to give the impression of a “window” on a virtual world.

Registration of 3D models from MRI
using geometric constraints for the
anatomical modeling of soft tissues

Licence Thesis

Professor: Nadia Magnenat-Thalmann
Assistant: Benjamin Gilles

MIRALab’s contribution in project Co-Me consists in developing a tool for the anatomical and functional modeling of the hip articulation from MRI (Imagery by Magnetic Resonance). In order to obtain precise models of the patient’s organs, this project proposes to use generic models (ligaments, cartilages, labrum) and to adapt them to the patients, by using two types of data: data resulting from the images of the patient (segmentation of characteristic zones) and geometrical constraints resulting from anatomical and topological properties.