A face recognition system based on a Kinect sensor and Windows Azure cloud technology

• Main Authors: Dobrea, Dan-Marius, Maxim, Daniel, Ceparu, Stefan
• Format: Conference Proceeding
• Language: English
• Subjects: Embedded systems; Face; Face detection; Face recognition; Neural networks; Training
• DOI: 10.1109/ISSCS.2013.6651227

The aim of this paper is to build a system for human detection based on facial recognition. The state-of-the-art face recognition algorithms obtain high recognition rates base on demanding costs - computational, energy and memory. The use of these classical algorithms on an embedded system cannot achieve such performances due to the existing constrains: computational power and memory. Our objective is to develop a cheap, real time embedded system able to recognize faces without any compromise on system's accuracy. The system is designed for automotive industry, smart house application and security systems. To achieve superior performance (higher recognition rates) in real time, an optimum combination of new technologies was used for detection and classification of faces. The face detection system uses skeletal-tracking feature of Microsoft Kinect sensor. The face recognition, more precisely - the training of neural network, the most computing-intensive part of the software, is achieved based on the Windows Azures cloud technology.