Real time recognition of 3D gestures in mobile devices

Abstract

Gesture-based user interaction is increasingly relevant today as the use of personal computing devices becomes widespread. Smartphones have several inbuilt sensors like accelerometer, orientation sensor and gyroscope, which are able to provide data on motion of the device in 3D space. This paper proposes a mechanism for real-time recognition of 3D gestures using sensors in mobile devices. 3D gestures are space-drawn gestures computed from 3-axial accelerometer readings. The algorithms discussed in this paper include single value decomposition, dynamic time warping and Mahalanobis distance.