The devil is in the details: A synthesis of psychology of facial perception and its applications in forensic facial reconstruction

Craniofacial identification standards: A review of reliability, reproducibility, and implementation

There are numerous anatomical and anthropometrical standards that can be utilised for craniofacial analysis and identification. These standards originate from a wide variety of sources, such as orthodontic, maxillofacial, surgical, anatomical, anthropological and forensic literature, and numerous media have been employed to collect data from living and deceased subjects. With the development of clinical imaging and the enhanced technology associated with this field, multiple methods of data collection have become accessible, including Computed Tomography, Cone-Beam Computed Tomography, Magnetic Resonance Imaging, Radiographs, Three-dimensional Scanning, Photogrammetry and Ultrasound, alongside the more traditional in vivo methods, such as palpation and direct measurement, and cadaveric human dissection. Practitioners often struggle to identify the most appropriate standards and research results are frequently inconsistent adding to the confusion. This paper aims to clarify how practitioners can choose optimal standards, which standards are the most reliable and when to apply these standards for craniofacial identification. This paper describes the advantages and disadvantages of each mode of data collection and collates published research to review standards across different populations for each facial feature. This paper does not aim to be a practical instruction paper; since this field encompasses a wide range of 2D and 3D approaches (e.g., clay sculpture, sketch, automated, computer-modelling), the implementation of these standards is left to the individual practitioner.

The Doppelgänger effect? A comparative study of forensic facial depiction methods

This study attempted to assess the reproducibility of 2D and 3D forensic methods for facial depiction from skeletal remains (2D sketch, 3D manual, 3D automated, 3D computer-assisted). In a blind study, thirteen practitioners produced fourteen facial depictions, using the same skull model derived from CT data of a living donor, a biological profile and relevant soft tissue data. The facial depictions were compared to the donor subject using three different evaluation methods: 3D geometric, 2D face recognition ranking and familiar resemblance ratings. Five of the 3D facial depictions (all 3D methods) demonstrated a deviation error within ± 2 mm for ≥ 50% of the total face surface. Overall, no single 3D method (manual, computer assisted, automated) produced consistently high results across all three evaluations. 2D comparisons with a facial photograph of the donor were carried out for all the 2D and 3D facial depictions using four freely available face recognition algorithms (Toolpie; Photomyne; Face ++; Amazon). The 2D sketch method produced the highest ranked matches to the donor photograph, with overall ranking in the top six. Only one 3D facial depiction was ranked highly in both the 3D geometric and 2D face recognition comparisons. The majority (67%) of the facial depictions were rated as limited or moderate resemblance by the familiar examiner. Only one 2D facial depiction was rated as strong resemblance, whilst two 2D sketches and two 3D facial depictions were rated as good resemblances by the familiar examiner. The four most geometrically accurate 3D facial depictions were only rated as limited or moderate resemblance to the donor by the familiar examiner. The results suggest that where a consistent facial depiction method is utilised, we can expect relatively consistent metric reliability between practitioners. However, presentation standards for practitioners would greatly enhance the possibility of recognition in forensic scenarios.

A revised nose tip shape validation method for facial reconstruction based on CT data from a modern German population

Several methods aid with reconstructing features of the human nose, including angle, projection and width, but only one study by Davy-Jow et al. (2012) has focused on nose tip shape. The main finding was that the shape of the nasal bridge is consistent with the shape of the nose tip. The study also theorised that the method would not be suitable for snub (upturned) noses. Although promising, further investigation with a larger sample of different origin would be of benefit. In addition, grouping samples into upturned, horizontal and downturned nose tips could reveal the need for a difference in the applied method. The approach has been recreated with a larger sample size (N = 103 versus N = 25) derived from a modern German population. Based on soft tissue models, the individuals were firstly grouped into three categories; upturned, horizontal, and downturned noses. Computed Tomography (CT) data allowed the simultaneous visualisation of both skull and (semi-transparent) facial surfaces. Each head was viewed frontally in the Frankfurt Horizontal Plane (FHP), and then tilted back until the nasal tip superimposed the nasal bridge, with the angle of tilt measured from the FHP. The results show that the angle of tilt is significantly different for upturned, horizontal, and downturned noses, but that it can be equally applied to all three groups. The mean angle was 44° for upturned noses, 51° for horizontal, and 56° for downturned. Error studies suggest a very high accuracy and repeatability with intra-class correlation coefficients of 0.991 (inter-observer error) and 0.972 (intra-observer error) respectively.

Facial reconstruction: Validation of the Tedeschi-Oliveira method for estimating the pronasale point in Brazilians

Forensic facial reconstruction aids in the process of human identification by facial recognition. The nose plays an important role in this process; however, its soft tissues wither away during cadaveric decomposition. There are few studies in the area of facial recognition of the Brazilian population, especially concerning Brazilian nasal prediction guides. This study aimed to test the hypothesis that the angle formed by the tangents drawn from the rhinion and prosthion points (nose tip) is equal to 90º, as proposed by the pioneer study by Tedeschi-Oliveira et al. (2016). It is important to highlight that this is the only method to date developed to predict the nasal tip in Brazilians, and has not yet been tested in other populations. Images of computed tomography scans of 228 individuals (171 females and 57 males) were screened according to the same criteria used by the authors of the primary study. The images were analyzed using the Horos® program, version 1.1.7 - 64 bit. The mean angle studied was about 90º, and any difference between this mean and the real angle measured did not significantly compromise the accuracy of the nose tip location. The findings tested in the present study corroborate the hypothesis investigated by Tedeschi-Oliveira et. al. Therefore, we suggest that the method be used in forensic practice to estimate nose tips in Brazilians.

Three-Dimensional Prediction of the Nose for Facial Approximation in a Thai Population

Most of the previous studies about nose prediction were concentrated only some landmarks of the nose. This study aimed to generate prediction equations for ten landmarks of the nose in the midline and alar regions for forensic facial approximation. The six midline landmarks were the sellion, nasion-pronasale posterior, nasion-pronasale anterior, pronasale, nasal drop, and subnasale. The four landmarks of the alar region were the alare, superior alar groove, posterior alar groove, and inferior alar groove. We used the skull landmarks in the nasal, zygomatic, and maxillary bone to predict the landmarks of the nose. Cone-beam computed tomography scans of 108 Thai subjects with normal BMI and age ranging from 21.0 to 50.9 years were obtained in a sitting position. The data were converted into three-dimensional (3-D) images of the skull and face. The Cartesian coordinates of the landmarks of the skull and nose were used to formulate the multiple regression equations. The formulated equations were tested in 24 new subjects. The mean differences in the predicted midline landmarks varied between -0.4 mm and 0.5 mm, whereas those for bilateral landmarks varied between -1.0 mm and 1.4 mm. In conclusion, the prediction equations formulated here will be beneficial for facial approximation of the nose in a Thai population.

Forensic Facial Approximation: Study of the Nose in Brazilian Subjects

Forensic facial approximation is an auxiliary method for human identification and allows facial recognition. The midface, that includes the nose, is vital for the recognition of a familiar face. The purpose of this study was to set hard tissue parameters to estimate nasal width, to test the method to estimate nasal width of Brazilians, and to analyze the relationship between nasal profile and facial type. A total of 246 computed tomography scans (183 females and 63 males) of adults were analyzed in Horos. Bone tissue measurements and facial type classification were performed on the skull scan. Nasal profile morphology was accessed through the tool 3D surface rendering. There was a difference around 3 mm from real to predicted nose through the method to estimate nasal width in Brazilians. So, the method may be used in forensic practice. Straight nose was associated with long face type.

Recognition of computerized facial approximations by familiar assessors

Studies testing the effectiveness of facial approximations typically involve groups of participants who are unfamiliar with the approximated individual(s). This limitation requires the use of photograph arrays including a picture of the subject for comparison to the facial approximation. While this practice is often necessary due to the difficulty in obtaining a group of assessors who are familiar with the approximated subject, it may not accurately simulate the thought process of the target audience (friends and family members) in comparing a mental image of the approximated subject to the facial approximation. As part of a larger process to evaluate the effectiveness and best implementation of the ReFace facial approximation software program, the rare opportunity arose to conduct a recognition study using assessors who were personally acquainted with the subjects of the approximations. ReFace facial approximations were generated based on preexisting medical scans, and co-workers of the scan donors were tested on whether they could accurately pick out the approximation of their colleague from arrays of facial approximations. Results from the study demonstrated an overall poor recognition performance (i.e., where a single choice within a pool is not enforced) for individuals who were familiar with the approximated subjects. Out of 220 recognition tests only 10.5% resulted in the assessor selecting the correct approximation (or correctly choosing not to make a selection when the array consisted only of foils), an outcome that was not significantly different from the 9% random chance rate. When allowed to select multiple approximations the assessors felt resembled the target individual, the overall sensitivity for ReFace approximations was 16.0% and the overall specificity was 81.8%. These results differ markedly from the results of a previous study using assessors who were unfamiliar with the approximated subjects. Some possible explanations for this disparity in performance were examined, and it was ultimately concluded that ReFace facial approximations may have limited effectiveness if used in the traditional way. However, some promising alternative uses are explored that may expand the utility of facial approximations for aiding in the identification of unknown human remains.

The unfamiliar face effect on forensic craniofacial reconstruction and recognition

Previous research into the reliability of forensic craniofacial reconstruction (CFR) has focused primarily on the accuracy of reconstructed faces from European or African ancestry skulls. Moreover, the recognition of CFR in relation to the experience and ancestry of the practitioners and the assessors has not been previously considered. The cross-race effect is a recognised phenomenon in psychology studies, where familiar ancestry faces are recognised more readily than unfamiliar ancestry faces, but there is a paucity of research addressing the relationship between the accuracy of reconstructed faces and the familiarity with this ancestry by the practitioners/assessors. The aims of this research were to investigate whether 'unfamiliar-race effect' has any influence on the accuracy of CFR and to evaluate how much the correct recognition rate of CFR is affected by the cross-race effect. Eight CFRs from three ancestry groups were produced by experienced practitioners in order to explore the aims. The results demonstrated that practitioners produced more recognisable CFRs using skulls from a familiar ancestry than skulls from unfamiliar ancestries.

Forensic facial approximation assessment: can application of different average facial tissue depth data facilitate recognition and establish acceptable level of resemblance?

Facial soft tissue thicknesses (FSTT) are important guidelines for modeling faces from skull. Amid so many FSTT data, Forensic artists have to make a subjective choice of a dataset that best meets their needs. This study investigated the performance of four FSTT datasets in the recognition and resemblance of Brazilian living individuals and the performance of assessors in recognizing people, according to sex and knowledge on Human Anatomy and Forensic Dentistry. Sixteen manual facial approximations (FAs) were constructed using three-dimensional (3D) prototypes of skulls (targets). The American method was chosen for the construction of the faces. One hundred and twenty participants evaluated all FAs by means of recognition and resemblance tests. This study showed higher proportions of recognition by FAs conducted with FSTT data from cadavers compared with those conducted with medical imaging data. Targets were also considered more similar to FAs conducted with FSTT data from cadavers. Nose and face shape, respectively, were considered the most similar regions to targets. The sex of assessors (male and female) and the knowledge on Human Anatomy and Forensic Dentistry did not play a determinant role to reach greater recognition rates. It was possible to conclude that FSTT data obtained from imaging may not facilitate recognition and establish acceptable level of resemblance. Grouping FSTT data by regions of the face, as proposed in this paper, may contribute to more accurate FAs.

Correlation between average tissue depth data and quantitative accuracy of forensic craniofacial reconstructions measured by geometric surface comparison method

Accuracy is the most important factor supporting the reliability of forensic facial reconstruction (FFR) comparing to the corresponding actual face. A number of methods have been employed to evaluate objective accuracy of FFR. Recently, it has been attempted that the degree of resemblance between computer-generated FFR and actual face is measured by geometric surface comparison method. In this study, three FFRs were produced employing live adult Korean subjects and three-dimensional computerized modeling software. The deviations of the facial surfaces between the FFR and the head scan CT of the corresponding subject were analyzed in reverse modeling software. The results were compared with those from a previous study which applied the same methodology as this study except average facial soft tissue depth dataset. Three FFRs of this study that applied updated dataset demonstrated lesser deviation errors between the facial surfaces of the FFR and corresponding subject than those from the previous study. The results proposed that appropriate average tissue depth data are important to increase quantitative accuracy of FFR.