Memorizing facial identity, expression and orientation

Abstract

Recognition for facial identity, expression, and orientation was investigated in a successive face comparison task. Subjects were required to make sarne/different judgements about pairs of face photographs that could differ in any one of these respects. Overall recognition performance for identity alterations was superior to that for expression and orientation changes. After a short retention interval (1 s) there was no difference between recognition accuracy for different responses to identity, expression and orientation alterations, but after a long delay (20 s) some expression and orientation information was forgotten while accuracy for identity judgements remained unchanged. Subjects could remember some expression and orientation information over a 20 s period, but memory for these dynamic attributes was less durable than identity memory.Possibly the most important property of a face is its identity. People are highly skilled at recognizing facial identity (e.g. Gombrich, 1972, p. 9). Moreover, facial memory is remarkably accurate after extremely long retention intervals (e.g. stretching over several decades (Bahrick et af., 1975)). Other properties of a face, for example the pose orientation and expression, are integral parts of the stimulus. However, although these aspects of a face may be perceived, a parsimonious recognition system might be able to extract and store identity information and neglect the less salient dynamic properties.The purpose of the present research was to investigate perception and memory for facial identity, expression, and pose orientation. The aim was to assess whether expression and orientation are perceived as accurately as identity and to determine whether the time course of memory is similar for these three attributes. It was also of interest to discover how expression and orientation interact with identity. in face recognition suggest that it is not a stimulus attribute that greatly affects facial memory. Laughery et al. (1 971) reported no significant differences in recognition accuracy between searching for a target face -originally presented in four candid views -in a series of photographs of front view faces, left profiles, right portrait views, and left portrait views. Davies et al.(1 978) demonstrated that identification accuracy was similar for full-face photographs and f profile photographs. Furthermore, altering the orientation of the faces from full face to f profile (and vice versa) between study and test was not detrimental to recognition accuracy.These data do not necessarily imply that subjects did not encode pose position or were unable to remember it. In fact, it has been demonstrated empirically that orientation changes between study and test phases of face recognition tasks can, in some cases, have adverse affects on recognition accuracy, which suggests that some orientation information can be stored. For example, Patterson & Baddeley (1977), like Davies et al., showed that after studying face photographs in a full-front view, a change to 3 profile had no si...

Encoding and storage effects in 7‐year‐olds' and 10‐year‐olds' memory for faces

“…Analogously, older children's performance may be superior to younger children's at immediate test because they initially encode more information in a given amount of time, although this is only retained on a short-term basis. It has been shown that adults encode certain facial feature information (relating to orientation and expression) which is only retained in memory for a limited period (Walker-Smith, 1980). Memory for facial characteristics may operate at several storage levels, information relating to expression and communication may be relatively transient compared with identity information which may be stored more permanently.…”

The impact of binge drinking on emotion recognition

“…The faces were learned and unlearned, changed viewing angle (frontal and three-quarter angle) and changed expression (smiling to unsmiling) or changed in both angle and expression. In agreement with previous research (Walker-Smith, 1980), Bruce found that unchanged faces were recognised more quickly and accurately than those with a change in angle or expression, which in turn were recognised more accurately than faces with changes in both. In the same study (Bruce, 1982) a second trial was conducted using familiar and unfamiliar faces with similar results to the first trial with the addition that unfamiliar faces were recognised more slowly and less accurately when a change had occurred, whereas familiar faces were recognised with the same accuracy as previously but at a slower rate.…”

Effects of pose, pose change, and delay on face recognition performance.

“…In addition, there is some indication that any effects due to pose and pose change might not remain stable over time. Although no investigators have studied pose and pose change after long delays, Walker-Smith (1980), using a "same-different" judgment task, found that information relevant to the discrimination of different facial poses declined during a very brief (20-sec) delay. Although the decline over a 20-sec period is not directly relevant to longer delays, it suggests that differences in performance between subjects who see different poses might drop out with delay.…”