Factors Limiting Face Matching at Passport Control and in Police Investigations

Face matching entails a comparison between two faces that are unfamiliar to an observer, who must then decide whether these depict the same person or different people. Despite the ubiquity of face matching in practical settings, such as passport control and police investigations, laboratory research has established that this task is highly error-prone, and that many of these errors derive from visual characteristics of to-be-compared face stimuli. Such characteristics include factors such as image quality, lighting, and natural changes in personal appearance, which influence the visual correspondence between face stimuli. In this chapter, factors that are likely to limit face-matching accuracy in real-world settings are reviewed, with the aim of providing insight into how these influence the accuracy of this process and how subsequent errors may be mitigated.

Epilogue

This book has provided an overview of the latest developments and current understanding of forensic/unfamiliar face matching. Throughout this book, the scientific study of face matching has been contextualized through two important applied tasks—facial comparison at passport control and in police settings. These applied tasks are now considered in turn to draw together content from different chapters and highlight some of the key insights emerging from this book....

Person Identification at Airports During Passport Control

This chapter provides insight into the development of internationally interoperable standards for passport and travel documents, and technological advances to facilitate fast and effective processing of passenger information for comparison against watchlists at borders and airports. The demands posed to border controls are considered and how these have changed over time, with particular emphasis on the modern-day threat of identity impostors and the development of facial recognition technology at Automated Border Control points. The continuing central role of human personnel for verifying the identity and nationality of travellers is described, with consideration of personnel selection, training, and performance monitoring.

Can Face Identification Ability Be Trained?

Is it possible to train face identification ability? To answer this question, we review the literature on professional training for facial image comparison practitioners, and the broader psychology literature on training for prosopagnosia patients and the general population. Our review of these literatures finds very little evidence that training can improve face recognition or face-matching accuracy. However, one group of specialist practitioners—facial examiners—throw this conclusion into question. Facial examiners consistently outperform novices and show qualitative differences in how they perform matching tasks, suggesting they have acquired their expertise via professional training. To account for these findings, we propose that there are two routes to expertise in face identification: The core face recognition system, which cannot be trained, and an alternative feature-based route, which can be trained. We encourage academics and practitioners to work together to identify effective and efficient training methods to teach facial image comparison practitioners to extract maximal identity information from facial features.

Forensic Face Matching

This chapter considers the use of forensic face matching as evidence in criminal trials. It provides an account of the evidential rules and principles that govern the admissibility of this form of evidence in common law jurisdictions—namely, the requirement of relevance, the prohibition on evidence of opinion, and the exception to that prohibition for those who possess expertise. It explains that notwithstanding differences in approach to the reception of expert evidence across jurisdictions, and despite persistent and unresolved concerns about its reliability, evidence of image-to-image comparison will be admitted. In contrast, because person-to-image comparison is a task that can usually be undertaken in the courtroom by the fact-finder, the evidence of witnesses who have engaged in this kind of comparison will usually be excluded on the grounds that it does not satisfy the fundamental requirement of relevance.

Realistic Masks in the Real World

Security and crime prevention often rely on facial appearance to connect individuals to behaviours. Hyper-realistic face masks can potentially frustrate this connection by allowing the wearer to look like someone else. In this chapter, we review the evidence that hyper-realistic masks are truly realistic, in the sense that they are accepted as real faces. We begin by outlining relevant experimental studies of face identification and disguise. We then tabulate all criminal cases known to involve hyper-realistic face masks (41 cases between 2009 and 2019). Experimental tests suggest that failures to detect such masks can be attributed to the realism of the masks, without invoking inattention or incompetence on the part of observers. We end with eight proposals for improving mask detection, encompassing training, personnel selection, and machine vision. If the misuse of hyper-realistic masks becomes widespread, our inability to detect them will compromise face recognition infrastructure.

Steps Towards a Cognitive Theory of Unfamiliar Face Matching

The visual comparison of unfamiliar faces—or ‘face matching’—is utilized widely for person identification in applied settings and has generated substantial research interest in psychology, but a cognitive theory to explain how observers perform this task does not exist. This chapter outlines issues of importance to support the development of a cognitive account of unfamiliar face matching. Characteristics of the face, such as within-person variability and between-person similarity in appearance, are considered as the visual input upon which identification must build. The cognitive mechanisms that observers may bring to bear on faces during identity comparison are analysed, focusing on attention, perception, evaluation, and decision processes, including sources of individual differences at each of these stages. Finally, the role of different experimental and occupational contexts in understanding face matching and for optimizing theory development is discussed.

Forensic Face Matching

Forensic face matching evidence has been presented in UK courts for over 30 years to provide crucial identification evidence in criminal investigations. To be admissible as evidence in UK courts, face matching must be conducted by a suitably qualified expert using scientifically validated procedures. Contrary to this notion, however, the field has been largely self-regulated, with little empirical investigation into the validity of face matching procedures and extensive criticism of forensic face matching in the scientific literature. Practitioner working groups are now addressing these criticisms and standardizing working practices, but further effort is required to ensure that the procedures used for forensic face matching are reliable and the limitations known. This chapter provides a critical analysis of the forensic face matching procedures used in the UK and internationally by forensic face examiners, alongside studies and case examples that have challenged and tested the reliability and accuracy of these procedures.

Individual Differences Between Observers in Face Matching

Increasing evidence suggests vast individual differences in face-matching performance in both lay perceivers and relevant professionals. However, the field is hampered by a paucity of psychometric-standard assessment tasks. This chapter reviews the current evidence supporting individual differences in face matching, in light of the reliability of the available tools. The potential underpinnings of these individual differences are reviewed, alongside the overlap between different continua of face-processing skills. Given that so-called super-recognizers may offer a powerful security tool in relevant settings, the identification and deployment of these individuals are critically discussed. In particular, the reliability and appropriateness of current terminology and assessment tools are considered, together with potential limitations in the performance of even the strongest face matchers. While the current conceptualization of super-recognition can no doubt advance academic theory, this approach may not adequately identify the best individuals for real-world forensic face-matching tasks.

Understanding Professional Expertise in Unfamiliar Face Matching

Deciding whether or not two images are of the same unfamiliar face is an important task in many professions. These decisions are a critical part of modern identity verification processes with direct—and often profound—consequences for individual rights and the security of society. As a result, the public expect the people entrusted with these decisions to perform accurately. But do they? Here we review 29 published tests comparing face matching accuracy in professional and novice groups. Twelve of these tests show no significant differences between professional and novice groups, suggesting that merely performing the task in daily work is not sufficient to improve accuracy. However, specialist groups of facial examiners and police super-recognizers consistently outperform novices. Staff selection, mentorship, deliberate practice, motivation, feedback and training may all contribute to enhancing the performance of individuals working in these groups, and future research is necessary to delineate their relative contributions.