Learning Scoliosis Patterns using Anatomical Models and Motion Capture

Idiopathic scoliosis is a progressive disease with multiple forms that mainly affects young women during their growth. Diagnosis of scoliosis and its severity, at all stages of the disease, is key to effective treatment. Current methods rely on clinical surface analysis and measurement of spinal deformity angles in static X-ray images. A challenge today is to enhance current clinical analysis methods by predicting internal spine characteristics from surface measurements. Therefore, the objective of this thesis is to correlate these external measurements with the internal characteristics of the patient. Beyond static analysis, this research also aims to develop new algorithmic tools for analyzing scoliosis through dynamic motion capture data.

Directors:

• Jean-Sébastien Franco
• Sergi Pujades
• Aurélien Courvoisier
• François Faure
• Edmond Boyer

Raporteurs:

• Lennart Scheys
• Claudio Vergari

Examinators:

• Jocelyne Troccaz
• Grégory Chagnon