Abstract
At the access control of the security room, the many of methods of identifying individuals are based on the use of ID cards. But these methods have problems in the sense that persons must use their hands to input the cards into the card reader every time they go in and out. The authors have developed a method of identifying individuals by using the stereo image of the name plate which is usually clipped on each breast of persons in many offices. After obtaining the name plate image by the camera, it is necessary to calibrate three dimensional misalignments in the direction of the name plate, which were caused at the time images were obtained. Therefore the stereo-image is used for three dimensional calibration. The three dimensional calibration is carried out by transferring the name plate image to the image estimated that it would be obtained in the case of setting the name plate opposite to the camera accurately. The identification is accomplished by using a neural network. In this paper, the authors describe the results of the experiments which were performed to demonstrate the effectiveness of the proposed method.
