DefinePK

Abstract

Nowadays, prolonged sitting becomes bad ergonomic habit within office workers. While they are too focusing on works, they do not realize this habit can lead to unhealthy lifestyle. This project proposed an algorithm of monitoring system for automated prolonged sitting detection at office workplaces by using Kinect Camera. The skeleton joints obtained from real-time skeleton data of Kinect camera will be analyzed in selecting the appropriate skeleton joint coordinates that can represent sitting and standing activities. This was done using the boxplot technique, which can identify the discriminate features of both activities. After that, the prolonged sitting detection system was developed based on the selected skeleton joint. The developed system will give user the real-time feedback of how long they had been sitting and record the sitting time. Analysis of the result revealed that there are 7 out of 20 skeleton joint coordinates can discriminate between sitting and standing with 98.33% accuracy of the developed system.