Submitted in partial compliance with the requirements for the Master’s Degree in Technology: Chiropractic, Department of Chiropractic, Durban University of Technology, Durban, South Africa, 2016. / Background There has been a remarkable increase in the participation of sport for athletes with disabilities. Consequently, there have been many international studies on injuries in athletes which have shown a high prevalence in wheelchair basketball, largely attributed to the fast-paced, high intensity nature of the sport. This sport has grown worldwide including South Africa, however, very little research has been published on South African wheelchair basketball players and more research is, therefore, needed. Aim To determine the prevalence and profile of musculoskeletal pain in elite wheelchair basketball players of different point classifications in South Africa. Hypothesis 1: Upper extremity (including neck and back) pain is experienced more commonly in lower point classified wheelchair basketball players than in higher point classified players. Hypothesis 2: Lower extremity pain is experienced more commonly in higher point classified players than in lower point classified players Method This study was a quantitative, cross-sectional, questionnaire-based study. The questionnaire comprised of sub-sections on demographics and disability characteristics; activity levels pertaining to wheelchair basketball and other sport/physical activity; the prevalence of pain and the impact thereof on wheelchair basketball and/or activities of daily living. This questionnaire was administered to 48 wheelchair basketball players who were competing in the 2015 Supersport League. A response rate of 70% was decided as the lower limit cut-off for statistical power. Results Fourty-three participants responded yielding an 89.58% response rate. The mean age of participants was 33.3 (SD:9.5) years and the majority of participants (n=35) were male and African (n=29). Out of the 43 participants, 79.1% (n=34) used mobility devices, the majority (n=20) used wheelchairs. Most of the participants (n=41) played wheelchair basketball for more than five years and 32 participants did not participate in other sport. Almost half of the participants (n=25) experienced musculoskeletal pain in the last twelve months or at present, 75% of whom (n=12) visited a Physiotherapist for the pain. More than half of these participants (n=15; 60%) reported that the pain negatively affected their basketball performance. It was established that arm pain occurred frequently in lower point classified players (1.0-2.5 point players) and that hand and wrist pain was also more prevalent in lower point players than in higher point players. The prevalence of lower extremity pain was low and there was no statistically significant difference between higher and lower point classified players. Conclusions and Recommendations The finding that upper extremity pain occurred more frequently in lower point classified players was in keeping with the first hypothesis (the null hypothesis was, therefore, rejected). The second hypothesis was, however, rejected (and the null hypothesis was, therefore, accepted) as lower extremity pain did not occur more frequently in higher point classified players than in lower point classified players. The Eta scores may have been higher and may have shown a much larger than typical relationship between point classification and the prevalence of musculoskeletal pain had there been a larger sample size. Notwithstanding this limitation, it is a challenge to obtain a significantly larger sample size due to the nature and limited number of participants in this sport. More studies are warranted on this group of individuals, as a large number experienced pain which affected more than half of the participants’ performance in wheelchair basketball. These studies are important for the future success of the South African players and the sport in South Africa. / M

Deep learning has shown to be successful on the task of semantic segmentation of three-dimensional (3D) point clouds, which has many interesting use cases in areas such as autonomous driving and defense applications. A common type of sensor used for collecting 3D point cloud data is Light Detection and Ranging (LiDAR) sensors. In this thesis, a time-correlated single-photon counting (TCSPC) LiDAR is used, which produces very accurate measurements over long distances up to several kilometers. The dataset collected by the TCSPC LiDAR used in the thesis contains two classes, person and other, and it comes with several challenges due to it being limited in terms of size and variation, as well as being extremely class imbalanced. The thesis aims to identify, analyze, and evaluate state-of-the-art deep learning models for semantic segmentation of point clouds produced by the TCSPC sensor. This is achieved by investigating different loss functions, data variations, and data augmentation techniques for a selected state-of-the-art deep learning architecture. The results showed that loss functions tailored for extremely imbalanced datasets performed the best with regard to the metric mean intersection over union (mIoU). Furthermore, an improvement in mIoU could be observed when some combinations of data augmentation techniques were employed. In general, the performance of the models varied heavily, with some achieving promising results and others achieving much worse results.

Light Detection and Ranging (LiDAR) sensors have many different application areas, from revealing archaeological structures to aiding navigation of vehicles. However, it is challenging to interpret and fully use the vast amount of unstructured data that LiDARs collect. Automatic classification of LiDAR data would ease the utilization, whether it is for examining structures or aiding vehicles. In recent years, there have been many advances in deep learning for semantic segmentation of automotive LiDAR data, but there is less research on aerial LiDAR data. This thesis investigates the current state-of-the-art deep learning architectures, and how well they perform on LiDAR data acquired by an Unmanned Aerial Vehicle (UAV). It also investigates different training techniques for class imbalanced and limited datasets, which are common challenges for semantic segmentation networks. Lastly, this thesis investigates if pre-training can improve the performance of the models. The LiDAR scans were first projected to range images and then a fully convolutional semantic segmentation network was used. Three different training techniques were evaluated: weighted sampling, data augmentation, and grouping of classes. No improvement was observed by the weighted sampling, neither did grouping of classes have a substantial effect on the performance. Pre-training on the large public dataset SemanticKITTI resulted in a small performance improvement, but the data augmentation seemed to have the largest positive impact. The mIoU of the best model, which was trained with data augmentation, was 63.7% and it performed very well on the classes Ground, Vegetation, and Vehicle. The other classes in the UAV dataset, Person and Structure, had very little data and were challenging for most models to classify correctly. In general, the models trained on UAV data performed similarly as the state-of-the-art models trained on automotive data.