A method for face recognition using a cognitive memory system is described in this chapter. A photograph was given showing many faces. Background clutter was present in addition to background noise. The problem was to identify all of the faces. A separate photo of an individual was also available. A 20 × 20 pixel low resolution version was created from the individual photo. From this, a large number of low resolution images were created by translation, rotation, scaling, and varying brightness and contrast. The 20 × 20 images looked like the images of a person’s face, but with low resolution, it would be impossible to identify the person. A low resolution cognitive memory was trained with the low resolution images. The low resolution images served as templates. The templates scanned the multi-face photo to locate and identify human faces. High resolution 50 × 50 images were then made from the detected facial images. A larger number of images were created from the detected images by translation, rotation, scaling, etc. They were presented to a high resolution cognitive memory that had been trained with labeled single photos of the individuals. Recognition was accomplished.