Distance matching based gesture recognition for healthcare using Microsoft's Kinect sensor

Gesture recognition for healthcare aiding independent living of young and elderly individuals has been a focused area of research. The purpose of this work is to identify 14 distinct gestures of individuals using similarity matching algorithm. The scope of this proposed work covers occupational hazards arising from prolonged sitting in a particular posture in employees of various companies. Gesture recognition has been achieved using seven similarity measures. This simple and fast technique contributes to field of health care under static gesture recognition as an application of machine learning with a high accuracy of 94.29% in 3.83 millisecond using city-block distance. The results have been statistically validated using Friedman Test.