An evaluation of technical study and performance for real-time face detection using Web Real-Time Communication

Abstract

This paper proposed the technical study of real-time face detection system and also cover a key technology. In order to find the most appropriate factors, techniques, and algorithm by evaluating the performance that included connection speed, and effectiveness of tracking and detecting a human face in various conditions. WebRTC worked perform in any condition of device and platform independence. Furthermore, WebRTC could operate securely on the HTTPS protocol. In the case of the transferring ability of images, it relied on the connection speed; that LAN and WiFi are the most prefer for this best image quality. The results shown that Haar-like feature and CLM have significantly detected the face area over the web browser in almost light conditions. However, Haar-like has the better precision when operated on the low-speed processor. The real-time face detection system could be identified the person who walked through the capture device. In addition, this proposed system could be applied to any scope of the personal identification system.