Development of a scapula markerless pose estimation based on deep learning approaches

The Kinesiology Laboratory (https://www.unige.ch/medecine/kinesiology/) and the Biomechanics Laboratory (https://www.unige.ch/medecine/chiru/fr/b-lab-tests-robotises-avances-de-dispositifs-chirurgicaux/) of the Geneva University Hospitals and University of Geneva have recently initiated a new clinical activity consisting of a biomechanical analysis of the shoulder. This analysis is currently focused on the diagnosis and follow-up of patients selected for a reverse shoulder arthroplasty (e.g. in case of rotator cuff-deficient shoulders and/or arthritis of the shoulder joint). Both shoulder kinematics and muscular activity (surface electromyography) are including in this analysis.

Among other bony segments, the scapula remains challenging to track during a motor task. This is mainly due to soft tissue artefacts (STA), i.e. the relative motion bone-sensors caused by the biological tissues between the sensors and the bone that leads to inaccuracies in the estimation of rigid bone poses or kinematics. Commonly, optoelectronic cameras are used to track the 3D trajectory of a set of cutaneous reflective markers placed on the subject skin with respect to specific bony landmarks. However, it has been demonstrated that during arm elevation, STA of several landmarks can reach up to 15 mm, leading to huge errors in bone pose.

The internship is for MSc level students performing their 5/6 months final research project in 2023. The position is full-time at FCBG in Campus Biotech. Ideally, the internship would start between February and May. Small monetary compensation.