Gesture identification for system navigation in 3D scene

摘要

This paper describes the approach used to detect hand postures in computationally undemanding manner. The approach employs some features of the Microsoft Kinect sensor and applies simple geometric calculations and simplifications to effectively detect static hand gestures in real time and under changing lighting and background conditions. The results are further applicable to dynamic gesture recognition, and will be used as a partial result in EU FPVs HBB-Next project.