Investigations into the reduction of soft tissue artefacts using projection of markers and microwave imaging

Movement analysis is a widely applied clinical tool for the diagnosis of neurological and musculoskeletal pathologies and in the evaluation of surgical interventions. The clinical gold standard for movement analysis is skin-mounted marker-based systems, whose clinical usability in computing the underlying movement of bones are impeded by soft tissue artefacts (STA). STA are discrepancies in bone movement calculated from skin-mounted markers, and are caused by the interposition of soft tissues.

Multibody kinematic optimisation (MKO) methods are the most widely applied solution to reduce STA. However, the efficacy of MKO methods varies between subjects and investigated motions, with most MKO methods not validated on participants with higher body mass index (BMI) scores.

This thesis proposes a practical solution to reduce the deleterious effects of STA. Two novel marker projection schemes, wherein the markers are projected onto the bone surface, are proposed. The projection schemes are validated on a dataset containing both skin-mounted marker trajectories and reference kinematics, of participants with varying BMI scores performing a wide variety of movements. Additionally, a novel imaging modality for biomechanics, microwave imaging, is investigated to project the markers onto the bone surface during both static and dynamic motion. The feasibility of this application of microwave imaging is investigated using both human models and tissue-mimicking phantoms.

Our results indicate that the projection schemes reduce errors in rotations most affected by STA, and also improves the quality of computed kinematics for all subjects and investigated motions. Additionally, the location of the bone can be detected using microwave imaging and a wearable system, in both static and dynamic situations. Our findings underscore the efficacy and generalisability of our multidisciplinary solution to reduce the effects of STA on computed kinematics, and represent a potential solution to improve the clinical usability of skin-mounted marker-based data.