Skip to Main Content
In this paper is presented an algorithm to detect human faces using the orthogonal projection principle PCA (Principal Component Analysis) and a neural network classifier trained offline with a set of training images acquired from a webcam as well as from the CBLC and Manchester databases. The algorithm for the detection and tracking of faces will be implemented in a Visual C + + application, which will be tested on the mobile robot. The purpose of this paper is to implement an image processing algorithm in robotics thus making a human-robot interaction. The experimental results have been achieved on the PeopleBot, a two-wheeled differently driven mobile robot with a castor wheel for stability.