Real-time human feature acquisition and human tracking by omnidirectional image sensor

Recently, research into personal identification and the automation of crime-prevention systems or room access administration systems has gained prominence. It is seen as being desirable to identify many people at the same time without any specific focus. Thus, the use of an omnidirectional image sensor called Hyper Omni Vision is effective, since this image sensor can observe a 360 degree field of view. In this paper, we propose a real-time human feature acquisition and human face tracking method from an omnidirectional image sequence. The characteristic of this method is not only the tracking of faces but also the memorization of useful face images for personal identification.