English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Technological advancements have spurred the usage of machine learning basedapplications in sports science. Physiotherapists, sports coaches and athletesactively look to incorporate the latest technologies in order to furtherimprove performance and avoid injuries. While wearable sensors are verypopular, their use is hindered by constraints on battery power and sensorcalibration, especially for use cases which require multiple sensors to beplaced on the body. Hence, there is renewed interest in video-based datacapture and analysis for sports science. In this paper, we present theapplication of classifying S\&C exercises using video. We focus on the popularMilitary Press exercise, where the execution is captured with a video-camerausing a mobile device, such as a mobile phone, and the goal is to classify theexecution into different types. Since video recordings need a lot of storageand computation, this use case requires data reduction, while preserving theclassification accuracy and enabling fast prediction. To this end, we proposean approach named BodyMTS to turn video into time series by employing body posetracking, followed by training and prediction using multivariate time seriesclassifiers. We analyze the accuracy and robustness of BodyMTS and show that itis robust to different types of noise caused by either video quality or poseestimation factors. We compare BodyMTS to state-of-the-art deep learningmethods which classify human activity directly from videos and show thatBodyMTS achieves similar accuracy, but with reduced running time and modelengineering effort. Finally, we discuss some of the practical aspects ofemploying BodyMTS in this application in terms of accuracy and robustness underreduced data quality and size. We show that BodyMTS achieves an averageaccuracy of 87\%, which is significantly higher than the accuracy of humandomain experts.

Fast and Robust Video-Based Exercise Classification via Body Pose Tracking and Scalable Multivariate Time Series Classifiers