Abstract : Locomotor disorders comprehension is limited by the absence of dynamic 3D imaging technology. 3D imagers give access to accurate but static information about bones morphology. On the other hand, motion analysis systems provide dynamic measures. However, these measures may be distorted by the presence of soft tissues between the bones and the skin surface.
First, we study a new method for correcting soft tissues artifacts when using external markers for motion estimation. This method is based on a surfacic approach and aims to follow the deformation created by the scapula on the skin surface. A robust registration algorithm, called IMCP, is used. This algorithm, developed in previous works, has been modified and adapted to be more specific to the study context: motion analysis using external markers. The improvements concern post-processing so as to make the most of the information mutualization properties of the IMCP, a way to take into account the influence of the edges of the markers clusters, and finally the optimization of the processing time thanks to multi-threading developments. In a second time, a specific MRI protocol is developed in order to allow morpho-functional analysis. Moreover, articular coherence indicators are proposed for the glenohumeral joint. These indicators are adapted to the case of errors in motion estimation.
The results show that the use of a markers cluster covering all the scapula do not allow to follow the scapula print at skin surface. Thanks to simulation studies, two hypotheses are proposed to explain these results: the noise created by the soft tissues is too significant, and / or the available skin resolution is not sufficient. The relevance of proposing markerless analysis is so emphasized. Our morpho-functional study on the glenohumeral joint shows the significant influence of the motion analysis protocol on joint coherence during bones animation. The placement of the markers on the scapula and on the humerus as well as the choice of the motion estimation algorithm have a large influence on the bony structures motion estimation.