Performance of typical and superior face recognizers on a novel interactive face matching procedure

Abstract

Unfamiliar simultaneous face matching is error prone. Reducing incorrect identification decisions will positively benefit forensic and security contexts. The absence of viewindependent information in static images likely contributes to the difficulty of unfamiliar face matching. We tested whether a novel interactive viewing procedure that provides the user with 3D structural information as they rotate a facial image to different orientations would improve face matching accuracy. We tested the performance of 'typical' (Experiment 1) and 'superior' (Experiment 2) face recognizers, comparing their performance using high-quality (Experiment 3) and pixelated (Experiment 4) Facebook profile images. In each trial, participants responded whether two images featured the same person with one of these images being either a static face, a video providing orientation information, or an interactive image. Taken together, the results show that fluid orientation information and interactivity prompt shifts in criterion and support matching performance. Because typical and superior face recognizers both benefited from the structural information provided by the novel viewing procedures, our results point to qualitatively similar reliance on pictorial encoding in these groups. This also suggests that interactive viewing tools can be valuable in assisting face matching in highperforming practitioner groups.Photo-ID is necessary for identity verification in a range of settings, from crossing borders to buying age-restricted goods and accessing services. However, despite the wide use of photo-ID, simultaneous face matching is error-prone. In controlled laboratory and field studies, error rates between 10 and 30% are typically observed (Bruce et al., 1999;Megreya & Burton, 2006. False alarms (i.e., incorrectly assigning two people to the sameThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Active exploration of faces in police lineups increases discrimination accuracy.

“…Like all lineup identification procedures, interactive lineups consist of a “package” of components (e.g., active exploration, pose-reinstatement, movement, multiple viewing angles, viewing a face from ¾ angle), and accuracy differences across lineup procedures may stem from any or a combination or all of these components. Research should continue to isolate the causal mechanisms that underpin enhanced performance in interactive lineups (e.g., Colloff et al, 2021; Smith, Andrews, et al, 2021). In Experiment 2, we began on this path and examined the active exploration mechanism, specifically drawing on diagnostic-feature-detection theory, to make predictions about how two additional types of interactive lineups—in which the faces are presented simultaneously—could further boost identification accuracy.…”

Low prevalence match and mismatch detection in simultaneous face matching: Influence of face recognition ability and feature focus guidance

“…The combined results of Experiment 1 and 2 suggest that deployment of SRs to identify verification critical roles may have a positive impact on the identity verification workplace. Indeed, recent research suggests that SRs would be useful in policing and identity verification roles as they are less impacted by face occlusions (e.g., face masks and sunglasses, Noyes, Davis, Petrov, Gray, & Ritchie, 2021) whereas interactive image matching procedures developed to aid face matching enhance SR's performance even further (Smith et al, 2021). Although effect sizes were far smaller, the use of internal facial feature guidance scales might also provide an additional advantage, albeit further research is required to define which scales may be most appropriate to employ in different workplace environments.…”

A signal detection–based confidence–similarity model of face matching.

“…The UVSD can account for basic findings in the face-matching literature, such as the presence of individual differences in matching ability (Bate et al, 2021;Smith et al, 2021;White et al, 2014; This document is copyrighted by the American Psychological Association or one of its allied publishers.…”

“…The UVSD can account for basic findings in the face-matching literature, such as the presence of individual differences in matching ability (Bate et al, 2021; Smith et al, 2021; White et al, 2014; Wirth & Carbon, 2017), or differences in overall accuracy across different face-matching tests (Burton et al, 2010). Take for example, two of the most popular face-matching tests, the GFMT (Burton et al, 2010) and the KFMT (Fysh & Bindemann, 2018).…”

“…The UVSD model can provide valuable insights into the relations between confidence and accuracy (Peirce, 1877). Curiously enough, confidence ratings have been only sparsely used in face-matching studies (but see Smith et al, 2021). This is unfortunate given that confidence is a central determinant of judgment and decision making (Pleskac & Busemeyer, 2010) and has been recognized as such in related areas of eyewitness identification (Gronlund et al, 2014; Wixted & Mickes, 2014; Wixted & Wells, 2017).…”