The challenges in many forensic fields are related not only to often complex and unique casework but commonly also to difficulties in establishing best practice procedures, standard operating procedures (SOPs), and creating quality management strategies that comply with the rigorous scrutiny of the results and conclusions of forensic investigations.

Forensic disciplines that deal with human identification, such as forensic anthropology and forensic facial image identification (FFII), have common aspects in that they use anthropological knowledge of human variation in their analyses, have few structured educational programmes, and so far, have comparatively low numbers of accredited units. The latter is often explained by the uniqueness of casework concerning the establishment of identity from human skeletal remains or facial images. However, legal and social scrutiny has led to increased interest in accreditation as a means of ascertaining and documenting adherence to best practices in the forensic sciences.

Accreditation is defined as attestation, usually undertaken by a national accreditation body, that an assessed body (in this study, a facial image comparison (FIC) unit) meets the requirements set by harmonized standards and conveyed via a formal demonstration of its competence, impartiality, and consistent operation [1]. Currently, the most commonly followed standard requirement for forensic laboratory-based activities has been ISO/IEC 17025. Another standard, specific to forensic disciplines [2] and partly based on the ISO/IEC 17025, is currently under development and will cover various areas of forensically relevant activities, including the recording, collecting, and storage of items, analyses, interpretations, and reports.

FFII is based on the anthropological method of a one-to-one comparison of the morphological facial features of individuals in questioned images from surveillance cameras, private recordings, or identity documents with those in reference images such as mugshots [3, 4]. Similarities and differences in morphological facial features are identified and evaluated, usually following a feature list (e.g. E3149 Standard Guide for Facial Image Comparison Feature List for Morphological Analysis) and aim to answer whether the images represent the same or different person [3, 4].

Morphological feature comparison has been established as a highly accurate and reliable method for comparing questioned and reference facial images [5]. This is in contrast to facial recognition, which uses an intuitive, holistic approach and is based on the innate ability of humans, and the learned ability of machines, to recognize familiar faces [3, 4]. Using facial recognition rather than forensic identification has its place, but not for the legally binding identification of individuals [4, 6], since especially false positives may lead to adverse legal consequences for the accused [7].

FFII is undertaken by facial examiners, who can present their conclusions regarding identity as expert witnesses in court. While a few best practice recommendations are available [3, 4], sporadic reports indicate that there is little consistency in SOPs (when they exist) among FIC units [6]. This may be partly due to gaps in the scientific basis of FFII, as identified by Gibelli et al. [8], including a limited understanding of the biological variability of facial features, missing data on the reliability and validity of the methodology, and a lack of suitable databases that can be shared for training purposes. Quality management frameworks and accreditation requirements may help harmonize procedures in the FIC community.

This study aimed to acquire accreditation-relevant information regarding working conditions in FIC units by surveying facial examiners across Europe to identify potential factors affecting accreditation uptake in this forensic area.

Two electronic surveys were distributed:

Survey 2 was more comprehensive than Survey 1 and asked specific questions related to accreditation requirements. Two main topics addressed in the surveys and targeted in this paper were: (i) the educational background and training of facial examiners, and (ii) accreditation-relevant insights, including methods used. Both surveys are available in the Supplementary Materials (Note: questions not discussed in the results will be subject to future publications.) The responses were grouped for analysis according to the reported accreditation status of the FIC unit. Frequencies were calculated but no further statistical testing was performed due to the small sample size.

Thirty-four respondents (12 for Survey 1 and 22 for Survey 2) from 16 countries participated in the surveys (Table 1). The response rate in Survey 1 was 100%; however, the response rate in Survey 2 could not be exactly calculated because the number of facial examiners reached through the DIWG list was unknown. An assumption concerning the overall response rate for both surveys was made based on the 60 facial examiners from 41 institutions participating in the last DIWG proficiency test and, considering that only two-thirds of the institutions were European [9], a total response rate of approximately 80% was estimated.

All respondents were working for governmental organizations at the time of the surveys. Nine (26.5%) respondents reported that their FIC unit was accredited (to ISO/IEC 17025). One respondent was unsure whether their unit was accredited. (Their response was included in the nonaccredited group.)

Overall, 17 (50%) respondents had worked as facial examiners for at least 6 (maximum 20) years. The team size of FIC units varied from 3 (13.6%) to 4–6 (40.9%), to 7–10 members (45.5%). In total, 77.8% of respondents in accredited FIC units worked as facial examiners for ≥6 years, compared with 40% in nonaccredited units (Figure 1). Respondents from accredited FIC units reported a smaller team size (3–6 members) more often than respondents from nonaccredited units (83.3% vs. 43.8%; Figure 2).

In total, 24 (70.6%) respondents had an academic background, 6 (17.6%) had a law enforcement background, and 4 (11.8%) reported both backgrounds. All respondents, except for one, were trained on the job. One reported self-study as a response to how they acquired training in FIC. In accredited units, 55.6% had an academic and 44.4% had a law enforcement background, while in nonaccredited units, 76.0% had an academic background and 8.0% had a law enforcement background (16.0% reported both; Figure 3).

A word cloud (Figure 4) represents the distribution of fields of academic degrees held by the respondents, with the most common being anthropology, biology, and forensic science, followed by anatomy and law.

When asked about FIC training schemes (more than one option was possible), 23 (67.6%), 20 (58.8%), 16 (47.1%), and 9 (26.5%) respondents reported internal mentoring, external training courses, internal training courses, and external online courses, respectively. Five respondents reported a combination of all four schemes, while 10 (14.7%) (29.4%) were only trained internally within their institutions. In general, more respondents (77.8%) from accredited units reported internal mentoring as a training method than those from nonaccredited units (64.0%). Respondents from nonaccredited units indicated that their FIC training comprised internal (52.0%) and external in-person courses (64.0%) more commonly than the accredited respondents (33.3% and 44.4%, respectively, Figure 5).

Concerning areas of expertise, 97.1% of respondents reported having experience in image-to-image comparisons, 67.6% in image processing, 58.8% in supervision or mentoring, 52.9% in review 1:n, 47.1% in automated facial recognition systems (AFRSs), 47.1% in other biometrics, and 23.5% in person-to-image comparisons. Figure 6 shows the distribution of the areas of expertise among respondents from accredited and nonaccredited FIC units. More respondents from accredited units had experience in supervision/mentoring (77.8% vs. 52.0%) and image processing (77.8% vs. 64.0%), while more respondents from nonaccredited units had experience in review 1:n (56.0% vs. 44.4%), AFRS (52.0% vs. 33.3%), and other biometrics (56.0% vs. 22.2%).

All respondents from accredited units reported having SOPs in place (Figure 7) and participating in the annual DIWG proficiency test, compared with 72.0% and 92.0% of respondents from nonaccredited units, respectively. The uptake of other proficiency tests was reported relatively equally among respondents from accredited (66.7%) and nonaccredited units (62.5%); (Figure 8).

Intralaboratory validation was reported by 83.3% of respondents from accredited units vs. 18.8% from nonaccredited units. Annual intralaboratory validation was indicated by 33.3% of respondents from accredited units, compared with 6.3% from nonaccredited ones. A quarter of the respondents from the nonaccredited units were unsure whether they performed intralaboratory validations (Figure 9).

When asked about the methods used in the FIC unit (with more than one option possible), all respondents from the accredited units replied that they used standard methods compared with 84.0% from nonaccredited units. A small proportion (11.1%) of the respondents from accredited units also reported using modified methods, while respondents from the nonaccredited units indicated using modified (44.0%) and self-developed methods (24.0%) (Figure 10).

Respondents from accredited units reported method validation annually or at other preset intervals (83.3%) or when a new method was introduced (50.0%). Responses from non-accredited units were as follows: validation annually or at other preset intervals (12.5%), when a new method was introduced (18.8%), when a modified method was introduced (31.3%), and 43.8% of respondents were unsure when method validations were performed (Figure 11).

This paper presents the results of two surveys administered to European facial examiners and aimed to assess working environments and accreditation-relevant information in FIC practice. Responses from accredited and nonaccredited FIC units were compared.

All respondents worked for governmental organizations, with approximately a quarter reporting that their FIC unitwas accredited (all to ISO/IEC 17025). Half of the respondents had worked as facial examiners for 6 to 20 years. In the accredited units, more than three-quarters of the respondents worked as facial examiners for 6 or more years vs. 40% in nonaccredited units. Most FIC units (86%) had between 4 and 10 members. Almost twice as many respondents (83%) from accredited units reported smaller team sizes of 3 to 6 members compared with nonaccredited units (44%). Hence, FIC units with more experienced facial examiners working in smaller teams were more likely to be accredited. Smaller FIC units may, however, be part of larger teams, with accreditation being sought for multiple disciplines in one accreditation application.

In the following paragraphs, reworded extracts from ISO/IEC 17025: 2017 and best practices guidelines precede the discussion text to link accreditation requirements with the survey responses. In the accreditation process, nonconformities in any aspect of the forensic unit's work, including laboratory examinations, results, or expert witness testimony are assessed against its policies and procedures, as evidenced by documentation, processes, systems, records, and reports, among others. Laboratories are required to “document its procedures to the extent necessary to ensure the consistent application of its laboratory activities and the validity of the results” [10].

While the range of required competencies depends on the operational requirements of a facial examiner's employer and the type of casework undertaken, a list of recommended knowledge, ability, and awareness for facial examiners can be found in Appendix A of the Best Practice Manual for Facial Image Comparison [3]. Similarly, Facial Identification Scientific Working Group (FISWG) [11] states that facial examiners “have to draw on a large foundation of knowledge, skill, and ability to accurately reach their conclusions. Additionally, the articulation of the scientific and legal basis for the expression of conclusions for many forensic, intelligence, or law enforcement purposes requires an even more advanced level of training to include an expanded set of knowledge, skills, and abilities above the level of basic concepts”. This means that, apart from the functional competencies of mastering the technical aspects of FIC, facial examiners need core competencies, such as problem-solving and communication skills, professional integrity, and impeccable work ethics, which are essential for working within the forensic sciences. In addition, cross-functional competencies are required, such as knowledge of technical aspects of images, or a formal understanding of AFRS [11].

Most of the responding facial examiners in this study had an academic education. However, in accredited FIC units, there was nearly an equal proportion of respondents with academic and law enforcement backgrounds, while in non-accredited units, only 8% of the respondents had a law enforcement background and 16% had both an academic and law enforcement background. This suggests that facial examiners with a law enforcement background contribute competencies that are favorable for accreditation. This may result from facial image comparisons having been developed as a discipline within law enforcement [5]. However, other governmental organizations, such as immigration and border control services, have recently begun to employ FIC rather than purely facial recognition.

The most common fields of academic degrees held by the respondents were anthropology, biology, and forensic science, as morphological facial feature assessment is based on an anthropological methodology, and a biology-based knowledge of human variation and forensic work competencies are essential in FIC. Previous studies have shown that forensic anthropologists work in FFII in most European countries [12, 13]. An educational background in biological anthropology may provide a good understanding of the requirements for working in FFII. Although some aspects of FIC, including an in-depth understanding of the technical characteristics of images and the investigative/forensic nature of assessments, may need to be acquired through additional training. Consequently, all respondents, except one, reported having received on-the-job training. Approximately a quarter of respondents reported external online courses and 60% mentioned external in-person courses as part of their training. However, internal mentoring was the most reported training type (68% overall) with 78% of respondents in accredited units receiving it. This may also reflect the fact that respondents from accredited units reported being employed for longer than those in nonaccredited units. Since the online and external training options for FIC have emerged only recently, their uptake may increase in the future. The variety of responses to our surveys indicated that training frameworks for facial examiners are not yet harmonized across Europe, complicating the formulation of job descriptions and the planning of initial and continuing training needs in the field.

Harmonized training may be especially important, as Towler et al. [14] showed that short professional training courses improve identification accuracy inconsistently, marginally, or not at all, despite trainees believing their performance had improved. Therefore, mentorship schemes (which were commonly reported by respondents) may be a valid training approach when appropriately documented and, if possible, validated. Moreover, FISWG [15] recommends that the on-the-job mentoring for future facial examiners should take at least 2 years.

The facial examiners who responded to our survey reported a broad spectrum of expertise, with nearly half listing experience with other biometrics, such as fingerprint comparisons and AFRS, and approximately 60% reporting experience with supervision or mentoring. Interestingly, review 1:n was reported as an expertise by only 53% of the respondents, showing a divide between the professional pathways of facial reviewers and facial examiners. A comparison between respondents from accredited and nonaccredited units showed that more than three-quarters of facial examiners in accredited units had experience with supervision/mentoring compared with approximately half of the facial examiners in nonaccredited units (more of whom reported experience in facial reviewing, AFRS, and other biometrics). The more common supervisory experience of respondents from accredited units may be related to their longer serving times. In contrast, respondents from nonaccredited units had more diverse roles and, thus, a potentially different pathway to the role of facial examiner, including review 1:n, AFRS, and particularly experience with other biometrics. The respondents from nonaccredited units also reported coming from larger teams than those from accredited units, which may allow for the diversification of roles and more options to acquire different types of expertise.

Proficiency testing is defined as evaluating participant performance against pre-established criteria via interlaboratory comparisons [16], which involve the same or similar items by two or more laboratories in accordance with predetermined conditions (ISO/IEC 17043, 3.4).

Interaboratory comparisons, the use of reference or quality control materials, and reviews of reported results are listed as some of the required monitoring tools. Although not explicitly stated in the ISO standard, annual intra- and interlaboratory comparisons (proficiency testing) are recommended [3], as they relate to the assessment of personnel competency and method validation, both of which need to be maintained and regularly monitored.

All respondents from accredited units reported having SOPs in place and participating in annual DIWG proficiency tests, compared with 72% and 92% from nonaccredited units, respectively. Approximately two-thirds of the respondents in accredited and nonaccredited units reported participation in external proficiency tests other than those organized by DIWG. Whether the respondents from nonaccredited units had no SOPs or no knowledge of them, in either case, it would be viewed as a nonconformity during an accreditation process, as personnel must be able to access any documentation that is applicable to their responsibilities.

Intralaboratory validations (or comparisons) were reported by 83% of respondents from accredited units compared with one-fifth of those from nonaccredited units. However, annual intralaboratory validation was indicated by only a third of the respondents from accredited units. Moreover, a quarter of the facial examiners from nonaccredited units were unsure whether intralaboratory comparisons were performed. Although blind intralaboratory testing can be part of validation schemes, personnel need to be informed about the results of such testing as part of their regular competence assessments.

Validation is defined as confirmation of a claim, through the provision of objective evidence, that the requirements for a specific intended use have been fulfilled [17]. Validation techniques can encompass calibration, evaluation of bias using reference materials, systematic assessment of factors influencing results, comparing results with other validated methods, interlaboratory comparisons, and/or evaluations of uncertainty.

According to ENFSI ([3], p. 13), “any method used to analyze and compare facial imagery should be validated by the laboratory to demonstrate the accuracy of the method and show that it produces reliable results” primarily because of the absence of empirical research in favor of any given method for FIC casework.

All respondents from accredited units indicated that they use standard methods for FIC, compared with 84% from nonaccredited units. Nearly half of the respondents from the nonaccredited units reported using modified methods, and a quarter reported using self-developed methods in their FIC casework. Of the respondents from accredited units, 83% reported annual method validation or at other preset intervals, and half when a new method was introduced. In contrast, only 13% of respondents from nonaccredited units indicated annual method validation or at other preset intervals. Approximately one-fifth mentioned validation when a new method was introduced, a third reported validation occurring when a modified method was introduced, and nearly half were unsure when (or whether) method validation was performed. Given the importance of validation for appropriate method application, the latter finding is especially concerning.

This survey ascertained the variety of tasks, working conditions, and experiences of European facial examiners. The diversity we found may create difficulties in conforming to accreditation requirements, including awareness of job descriptions, creation of SOPs that fit different tasks, familiarization of personnel with these various documents, and documentation of competence (i.e. sufficient training and knowledge) in several, often rapidly changing areas such as the current state of applicability of AFRS or statistical and evaluative frameworks. The growing demand for competence in FFII has thus far been accompanied by the limited availability of structured education programmes that can be undertaken, apart from on-the-job training and internal mentoring schemes.

A similar situation regarding working conditions can be found in forensic anthropology, where discipline-specific educational programmes are available in only a few European countries, and there is a constant demand to stay up-todate (e.g. regarding imaging technologies) while caseloads increase, particularly in areas of disaster victim identification and humanitarian action.

Time constraints stemming from high caseloads and the demands of supporting documentation for the accreditation process may hinder seeking out accreditation as confirmed by the low number of accredited FIC units. However, the responses from facial examiners from accredited FIC units suggest that understanding and investing in best practices, such as having SOPs in place, following standard methods, performing intra- and interlaboratory comparisons or proficiency testing, and conducting regular method validations are favorable to accreditation.

In principle, best practices primarily mirror accreditation requirements; as such, these are not only a means to an end (i.e. accreditation) but are beneficial in streamlining routine practice and helping to navigate the complexities of casework. Close cooperation and the exchange of expertise among forensic practitioners and quality managers may also facilitate accreditation uptake in the future.

First and foremost, we are grateful to all the survey participants for their time and insights. We thank Patrick DeSmet for distributing the survey via the DIWG emailing list.

Zuzana Obertová conceived of the study, participated in its design, helped draft the surveys, performed the statistical analysis, undertook the interpretation of results, and drafted the manuscript; Inga Siebke participated in the design of the study, drafted the surveys, and participated in the interpretation of results; Grit Schüler conceived of the study, participated in its design, helped draft the surveys, and participated in the interpretation of results. All authors contributed to the final text and approved it.

Participation in the surveys was voluntary and anonymous. The participants consented to the publication of the results. This study was exempt from an Ethics Committee approval because no medical data were collected, and completion of the survey was voluntary and anonymous.

The authors report that there are no competing interests to declare.

This research received no specific funding. However, the AFORE training was funded by the European Union's Internal Security Fund—Police [project 875746-AFORE-ISFP-2018-AG-IBA-ENFSI].