A robust viewing angle independent hand gesture recognition system

The use of hand gesture for Human Computer Interface (HCI) is very common nowadays with the help of smart devices such as smart TV, smart phone, etc. and in recent years vision-based gesture recognition system has received a great attention. The task of recognition in real time and with high accuracy is becoming the main priority. In gesture recognition system one commonly encountered problem is detecting the gesture when made from side angles. In this work, a solution to that problem is derived. In this paper, a hand gesture recognition system for multiple viewing angles using a single camera based on combination of Affine Transform and Discrete Fourier Transform (DFT) is presented. Experimental results show that the system can efficiently detect gestures made from multiple angles with an average recognition rate of 95.28% and 90.30% for gestures made at ±30° and ±45° respectively.