The Joint Accelerator Conferences Website (JACoW) is an international collaboration that publishes the proceedings of accelerator conferences held around the world.
TY - CONF AU - Bacher, R. ED - Cheng, Yung-Sen ED - Schaa, Volker RW ED - Chiu, Pei-Chen ED - Li, Lu ED - Liu, Yung-Hui ED - Petit-Jean-Genaz, Christine TI - Improving Web2cHMI Gesture Recognition Using Machine Learning J2 - Proc. of PCaPAC2018, Hsinchu, Taiwan, 16-19 October 2018 CY - Hsinchu, Taiwan T2 - International Workshop on Emerging Technologies and Scientific Facilities Controls T3 - 12 LA - english AB - Web2cHMI is multi-modal human-machine interface which seamlessly incorporates actions based on various interface modalities in a single API, including finger, hand and head gestures as well as spoken commands. The set of native gestures provided by off-the-shelf 2D- or 3D-interface devices such as the Myo gesture control armband can be enriched or extended by additional custom gestures. This paper discusses a particular method and its implementation in recognizing different finger, hand and head movements using supervised machine learning algorithms including a non-linear regression for feature extraction of the movement and a k-nearest neighbor method for movement classification using memorized training data. The method is capable of distinguishing between fast and slow, short and long, up and down, or right and left linear as well as clockwise and counterclockwise circular movements, which can then be associated with specific user interactions. PB - JACoW Publishing CP - Geneva, Switzerland SP - 148 EP - 150 KW - controls KW - toolkit KW - interface KW - extraction KW - TANGO DA - 2019/01 PY - 2019 SN - 978-3-95450-200-4 DO - DOI: 10.18429/JACoW-PCaPAC2018-THCB3 UR - http://jacow.org/pcapac2018/papers/thcb3.pdf ER -