Facial recognition involving face location in a down-sampled image

摘要

A facial recognition method and system comprises: capturing a first image at a first resolution from a video camera 104; down-sampling 154 the first image 110 to create a second image at a lower resolution 112; searching the second image to find faces, and identifying their coordinates 156; extracting facial images from the first, full-resolution, image using the coordinates derived from the second image 158; and submitting each facial image to a facial recognition program and receiving identity data in response thereto (242). The method may also include the use of a quality measure to discard some images. The method may also include prioritization and filtering steps that choose which captured images from a stream of images to process and when to process them when there are many images to be processed.