This study is an exploratory work into semantic segmentation of rowing images. Rowing is a highly technical sport, which is very suitable for automated analysis. However, not many systems are available for this yet, with the ones that are available using inertial sensors. Being ableto analyse (old) rowing footage could help coaches further improve their crew`s technique. This study aims totake a first step towards visual automated analysis of therowing stroke. In this paper, we retrained a pre-trained Deeplabv3+ model to segment rowers and their boats. The performance of the model was evaluated similarly to Microsoft`s COCO challenge, with the primary metric being the mean intersection over union and pitted against the performance of the pre-trained model. The results show an increase in performance of 14.5% in the primary metric when using the retrained model, even though a very limited amount of training was done. These results show that there is potential in using machine learning to create an automated video analysis system for application in rowing.
© Copyright 2020 Veröffentlicht von University of Twente. Alle Rechte vorbehalten.