Study and Development of Face Recognition for Home Robots

• Main Authors: Chia-Yun Wu, 吳佳運
• Other Authors: Kuo-Yang Tu
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/58422227986738947429

碩士 === 國立高雄第一科技大學 === 系統資訊與控制研究所 === 98 === This paper presents a suitable robot to use face recognition system that can detect real-time images of the face, and identify the detected face images. One system consists of three subsystems: human face detection, face tracking and face recognition. In the face detection system, using Haar features, integral imaging, Adaboost algorithm and the waterfall model. Through Haar features to determine whether someone within the region face the existence of eigenvalues. Method of using the integral image to speed up the calculation of Haar features. The purpose of the use of Adaboost algorithm arranged the way, choose a specific size and location of Haar features, making more efficient face detection. In the waterfall model approach, you can quickly filter out in line with face area. In the face tracking system, through Fuzzy control to simple and fast way to adjust the angle of the camera to track the face. In the face recognition system, using the Principal Component Analysis obtained Eigenfaces. Calculate the Euclidean distance, has been identified in the face image. This paper is designed for face recognition system that can instantly detect face and face recognition. Very suitable for dynamic behavior of the robot, you can easily reach home care, service, security and other functions.