Real-time optoelectronic morphological processor for human face recognition
Haisong Liu · Optical Engineering · 1998
Many commercial and law enforcement applications of face recognition need to be high-speed and real-time, such as passing through customs quickly while ensuring security. However, face recognition by using computers only is time-consuming due to the intensive calculation task. Recently optical implementations of real-time face recognition have attracted much attention. In this paper, a real-time optoelectronic morphological processor for face recognition is presented. It is based on original-complementary composite encoding hit-or-miss transformation, which combines the foreground and background of an image into a whole. One liquid-crystal display (LCD) panel is used as two realtime SLMs for both stored images and face images to be recognized, which are of 256x3256 pixels. A speed of 40 frames/s and four-channel recognition ability have been achieved. The experimental results have shown that the processor has an accuracy over 90% and error tolerance to rotation up to 8 deg, to noise disturbance up to 25%, and to image loss up to 40%.