Face recognition in a tactical environment

Recently, a face recognition algorithm has been proposed that has been designed to work in uncontrolled settings with non-cooperative subjects. This technique is particularly useful for access control in a tactical setting due to ease of use. The method offers secure access control while reducing the cognitive load on the warfighter so that he can concentrate on the mission. The method works with standard webcams and can be used for logon and screen lock capabilities. This method can be incorporated into a multi -factor solution where smart card and other biometric modalities can be added for enhanced security. This technology can also be applied to stand-off applications or surveillance type scenarios where the warfighter can capture a video stream from a local village and process it against a "watch list". The face recognition algorithm can recognize the individual under varying conditions that may be found in such an application such as changes in illumination, weather, facial position and expression and other degradations. One of the key features of this system is an automatic image quality and pose estimation scheme. This allows the capture of a video stream followed by automatic extraction of the key frame or frames that contain the best quality faces for recognition. This can be used to filter the video stream and transmit only the information necessary to identify the subject of interest to a central server that contains the watch list information in a secure environment. This is especially important for tactical network environments where bandwidth is at a premium and transmission of high bitrate video streams may not be feasible.