Title :
Event-driven body motion analysis for real-time gesture recognition
Author :
Kohn, Bernhard ; Belbachir, Ahmed Nabil ; Hahn, Thomas ; Kaufmann, Hannes
Author_Institution :
AIT Austrian Inst. of Technol. GmbH, Vienna, Austria
Abstract :
This paper presents an evaluation of spatio-temporal data generated by a dynamic stereo vision sensor in a highdimensional space (3D volume and time) for motion analysis and gesture recognition. In contrast to traditional frame-based (synchronous) stereo cameras, dynamic stereo vision sensors asynchronously generates events upon scene dynamics. Motion activities are intrinsically (on-chip) segmented by the sensor, such that activity, gesture recognition and tracking can be intuitively and efficiently performed. In this work, we investigated the applicability of this sensor for gesture recognition. We developed a machine learning method based on the Hidden Markow Model for training and automated classifications of gestures using the event data generated by the sensor. By training eight different activities (dance figures) with 15 persons we build up a library of 580 recorded activities. An average recognition rate of 97% has been reached.
Keywords :
gesture recognition; hidden Markov models; image classification; image motion analysis; image sensors; learning (artificial intelligence); object tracking; stereo image processing; 3D volume; automated gesture classification; average recognition rate; dance figure; dynamic stereo vision sensor; event data; event-driven body motion analysis; hidden Markov model; high-dimensional space; machine learning; motion activity; real-time gesture recognition; scene dynamics; spatio-temporal data; tracking; Detectors; Dynamics; Gesture recognition; Hidden Markov models; Stereo vision; Training; Vectors;
Conference_Titel :
Circuits and Systems (ISCAS), 2012 IEEE International Symposium on
Conference_Location :
Seoul
Print_ISBN :
978-1-4673-0218-0
DOI :
10.1109/ISCAS.2012.6272132