S.A. Suandi, S. Enokida, and T. Ejima (Japan)
Computer vision, tracking, template matching, facial fea tures.
In this paper, we introduce an extended technique in tem plate matching to track eyes and mouth in real-time. The technique makes use of a set of n correlation candidates from template matching. We first list all the candidates from each face model regions, and select the best candi dates based on two selective functions, Seyes and Smouth. These functions are for right-left eyes pair and eyes-mouth pair selection, respectively. We also introduce a novel tech nique in tracking framework, called feature selective(FS), where the system selects the features automatically so that it is feasible for multiple face types and conditions. Exper iments show that significant improvements in tracking pre cision can be achieved when the candidates are increased, and even improved better when FS is added. Results also reveal that an average tracking precision of 97% for all targets may be achieved using our proposed method. We also show that our method is capable to handle various face types and postures.
Important Links:
Go Back
