Multi-Frame Image Restoration for Face Recognition

Face recognition at a distance is a challenging and important law-enforcement surveillance problem, with low image resolution and blur contributing to the difficulties. We present a method for combining a sequence of video frames of a subject in order to create a restored image of the face with reduced blur. A generic Active Appearance Model of face shape and appearance is used for registration. By warping and averaging registered video frames, noise is reduced, allowing a Wiener filter to deblur the face to a greater degree than can be achieved on a single video frame. This process is theoretically justified and tested with real-world outdoor video using a PTZ camera and a commercial face recognition engine. Improvement is demonstrated for both face recognition and authentication.