Verification of face identities from images captured on video.

Abstract

Four experiments investigated matching of unfamiliar target faces taken from high-quality video against arrays of photographs. In Experiment 1, targets were present in 50% of arrays. Accuracy was poor and worsened when viewpoint and expression differed between target and array faces. In Experiment 2, targets were present in every array, but performance remained highly error prone. In Experiment 3, short video clips of the targets were shown and replayed as often as necessary, but performance levels were only slightly better than Experiment 2. Experiment 4 showed that matching was dominated by external face features. The results urge caution in the use of video images to identify people who have committed crimes. Superficial impressions of resemblance or dissimilarity between face images can be highly misleading.The human face provides the most reliable means of person identification available to the human eye (although fingerprints and iris patterns may prove more useful for automated identification; e.g., seeDaugman, 1998). Nonethe-

Super‐recognisers in Action: Evidence from Face‐matching and Face Memory Tasks

“…frontal view and top lighting) but dropped dramatically when pose or lighting type differed between pictures. These results were similar to those reported by Bruce et al (1999), who found that a change in viewpoint was particularly detrimental in unfamiliar face matching, where participants mainly used external features to make their judgements.…”

“…In an initial experiment, we investigated the performance of SRs and typical perceivers on a face‐matching task that has been well‐validated in previous research (Bindemann, Brown, Koyas & Russ 2012; Bruce et al, 1999; Megreya, Sandford & Burton, 2013). On each trial, participants were required to match a video still of an unfamiliar male face to an array containing 10 faces.…”

“…Yet, there are large individual differences in the ability to recognise (Bowles et al, 2009; Russell, Duchaine & Nakayama, 2009) and perceive (Megreya & Bindemann, 2013; Megreya & Burton, 2006) faces, and particular difficulties are associated with the processing of unfamiliar facial stimuli (see Hancock, Bruce & Burton, 2000 for a review). For instance, over 80 trials, Bruce et al (1999) asked participants to select a face from an array of 10 that matched the identity of a target face. Stimuli presented within the arrays were photographic images, whereas the target faces were still images acquired from video footage.…”

“…It has been widely established that, while identification of familiar faces from CCTV footage is highly reliable and accurate (Burton, 2013; Burton, Wilson, Cowan & Bruce, 1999), the task of identifying or even matching unfamiliar faces is characterised by a high rate of error (Bruce et al, 1999; Bruce, Henderson, Newman & Burton, 2001; Burton et al, 1999). Indeed, Burton et al (1999) demonstrated that the recognition of familiar faces, even from poor‐quality CCTV, is fast and effortless.…”

“…Such a scenario could occur when missing or wanted persons are spotted by police officers on patrol. No study to date has investigated the ability of SRs to match faces from video stills to photographs — a task that typical perceivers are remarkably poor at (Bruce et al, 1999). Similarly, while there is strong evidence that typical observers are poor at remembering unfamiliar faces (Burton et al, 1999), there are no data demonstrating the performance of SRs on this type of task.…”