Towards Automatic Modelling of Volleyball Players' Behavior for Analysis, Feedback and Hybrid Training

Fahim A Salim, Fasih Haider, Dees B.W. Postma, Robby Van Delden, Dennis Reidsma, Saturnino Luz, Bert Jan Van Beijnum

Research output: Contribution to journalArticlepeer-review

Abstract / Description of output

Automatic tagging of video recordings of sports matches and training sessions can be helpful to coaches and players, and provide access to structured data at a scale that would be unfeasible if one were to rely on manual tagging. Recognition of different actions forms an essential part of sports video tagging. In this paper, we employ machine learning techniques to automatically recognise specific types of volleyball actions (i.e. underhand serve, overhead pass, serve, forearm pass, one hand pass, smash and block which are manually annotated) during matches and training sessions (uncontrolled, in the wild data) based on motion data captured by inertial measurement unit (IMU) sensors strapped on the wrists of 8 female volleyball players. Analysis of the results suggests that all sensors in the IMU (i.e. magnetometer, accelerometer, barometer and gyroscope) contribute unique information in the classification of volleyball actions types. We demonstrate that while the accelerometer feature set provides better results than other sensors overall (i.e. gyroscope, magnetometer and barometer) feature fusion of the accelerometer, magnetometer and gyroscope provides the bests results (Unweighted Average Recall (UAR)= 67.87%, Unweighted Average Precision (UAP)= 68.68% and Kappa = 0.727), well above the chance level of 14.28%. Interestingly, it is also demonstrated that the dominant hand (UAR =61.45%, UAP= 65.41% and Kappa = 0.652) provides better
results than the non-dominant (UAR = 45.56%, UAP = 55.45 and Kappa = 0.553) hand. Apart from machine learning models, this paper also discusses a modular architecture for a system to automatically supplement video recording by detecting events of interests in volleyball matches and training sessions and to provide tailored and interactive multi-modal feedback by utilizing an html5/JavaScript application. A proof of concept prototype developed based on this architecture is also described.
Original languageEnglish
Pages (from-to)323–330
JournalJournal for the Measurement of Physical Behaviour
Volume3
Issue number4
DOIs
Publication statusPublished - 1 Dec 2020

Fingerprint

Dive into the research topics of 'Towards Automatic Modelling of Volleyball Players' Behavior for Analysis, Feedback and Hybrid Training'. Together they form a unique fingerprint.

Cite this