The proportion of a human face which needs to be visible for correct identification was investigated. Photographs of familiar public faces were presented under two conditions: (i) the facial photographs were original photographs with local information; and (ii) the facial photographs were blurred, with high spatial frequencies removed and with no local information. Subjects could scan the face as they wished through a moveable window generated by a TV system and had to identify the depicted person as quickly as possible. Window size was varied systematically. The main study showed that, in spite of including only its local features, a familiar facial photograph could be perfectly identified when the visual window covered one-fifth of the whole face, and that a global image of the whole face was obtained when nearly half of the whole face was processed simultaneously. The structure of the memory image of the face is discussed in the light of the results obtained.
Classroom practice for understanding pointers using learning support system for visualizing memory image and target domain world