Face Recognition Method Based on Fixed and PTZ Camera Control for Moving Humans
Seong-Hoon Kim, Jin‐Woo Choi · 2019
Face recognition methods have been generally used for authentication purposes, in which subjects position their faces in the camera field. Lately, face recognition has begun to be used for security applications at specific locations as well as for authentication. Face recognition for security purposes involves human and face detection based on the image created by a fixed camera. Consequently, the success rate for face recognition is low for low-resolution cameras. To address this problem, this study proposes a face recognition method in a multi-camera environment that includes both a fixed and a pan-tilt-zoom (PTZ) camera. The proposed method involves detecting humans within the fixed camera's field of view while tracking the detected person by adjusting the PTZ camera's pan and tilt to keep the person in view. By zooming simultaneously with tracking to obtain a high-resolution facial image, the proposed method outperformed conventional face recognition methods that use only a fixed camera.