Skip to Main Content
This paper presents an automatic real-time face recognition system installed on a person following robot. To efficiently identify the persons interacting with the robot, we apply the two-dimensional Haar wavelet transform (2D-HWT) to acquire the low-frequency data of a face image sequence. For the facilitation of the face recognition system, both the shortcomings of principal component analysis (PCA) method which cannot effectively distinguish from different classes and those of linear discriminant analysis (LDA) method which may find no inverse matrix are improved. Through vigorous evaluation, we then employ the discriminative common vectors (DCV) algorithm to setup the discriminative models of facial features received from different persons. Finally, the Euclidean distance is utilized to measure the similarity of a face image and a candidate person and decide the most probable person by the majority vote of ten successive recognition results from a face image sequence. Experimental results reveal that the face recognition rate is more than 93% in general situations and still reaches 80% in cluttered backgrounds.