Elsevier

Forensic Science International

Volume 246, January 2015, Pages 110-121
Forensic Science International

Review Article
Conceptual transitions in methods of skull-photo superimposition that impact the reliability of identification: A review

https://doi.org/10.1016/j.forsciint.2014.10.043Get rights and content

Highlights

  • Practitioners’ and researchers’ methods of skull-photo superimposition are reviewed.

  • Departure in using life size image, point in the ear plane, wipe mode and asymmetry.

  • Practitioner's adherence to these concepts shown as safe and better for comparison.

  • Reduction in the reliability may be due to the above variations in methodology.

  • Evaluating reliability of superimposition using practitioner's method is stressed.

Abstract

Establishing identification during skull-photo superimposition relies on correlating the salient morphological features of an unidentified skull with those of a face-image of a suspected dead individual using image overlay processes. Technical progression in the process of overlay has included the incorporation of video cameras, image-mixing devices and software that enables real-time vision-mixing. Conceptual transitions occur in the superimposition methods that involve ‘life-size’ images, that achieve orientation of the skull to the posture of the face in the photograph and that assess the extent of match. A recent report on the reliability of identification using the superimposition method adopted the currently prevalent methods and suggested an increased rate of failures when skulls were compared with related and unrelated face images. The reported reduction in the reliability of the superimposition method prompted a review of the transition in the concepts that are involved in skull-photo superimposition. The prevalent popular methods for visualizing the superimposed images at less than ‘life-size’, overlaying skull-face images by relying on the cranial and facial landmarks in the frontal plane when orienting the skull for matching and evaluating the match on a morphological basis by relying on mix-mode alone are the major departures in the methodology that may have reduced the identification reliability. The need to reassess the reliability of the method that incorporates the concepts which have been considered appropriate by the practitioners is stressed.

Introduction

Since the innovation of the skull-photo superimposition method reported by Professors Glaister and Brash in 1935 [1], the photographic process of image overlay has undergone significant technical progression [2], [3], [4], [5], [6], [7], [8], [9], [10] (Table 1). It has evolved into a real-time video process [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30] (Table 2) that has evoked considerable research aimed at automating the procedure using computers and software [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42] (Table 3). A much-cited report on the reliability of the video superimposition method by Austin-Smith and Maples [22] indicated a failure to exclude identifications of approximately 9% when comparing three skulls with 100 front-view photographs of living individuals. Recently, Gordon and Steyn [42] superimposed 40 digitized facial photographs with 3D digital images of 10 skulls each (a total of 400 superimpositions were done for each of the morphological and landmark methods) and observed that the failure to exclude (false positives) was 17.3% and 32% for the morphological and landmark methods, respectively. In addition, these authors [42] reported that the failure to include (false negatives) was 15% for the morphological method and 20% for the landmark method. The fact that skulls failed to match their related faces during superimposition in 15–20% instances appears counterintuitive. This finding raises troublesome questions because it has been relatively well accepted that the skull is the matrix of the living head … and face in life [43] and creates the architectural form of the head and provides the basic structure for the face [44], which is a concept shared by many other authors [20], [45], [46]. Furthermore, in view of the number of real-world cases in which the superimposition method has provided reliable results, the rate of failures reported by researchers [22], [42] appears inflated and has prompted a review of the concepts underlying the methodologies recommended by practitioners and those used by researchers to identify divergences, if any.

Section snippets

Practitioners’ and researchers’ methods: specific conceptual variances

The reported instances of applications of the superimposition method for identification purposes in routine case work include 25 in the CA Pound Laboratory, Florida [25], 52 in Japan [23], 108 in China [47], 71 in Hungary [48] and 251 in India during 1970–1989 [18], which increased to 1800 during 1990–2010 [49], together totaling 2307. A recent national level survey in India revealed that the number of superimposition-based court testimonies in the Tamil Nadu state during 2005–2009 was 200 [50]

‘Life-size’ face-image and skull orientation: the two critical requirements

Fundamentally, skull-photo superimposition is an image-mixing process that helps examine the appropriateness of the salient features in the image of the skull in question (that has been anatomically oriented to correspond to the posture in the face-image) for the missing individual's face-image that has been magnified appropriately; the classical example is that described by Glaister and Brash [1]. Subsequent authors describing the photographic process of overlaying transparencies [2], [4], [6]

On the concepts and methods relating to the use of ‘life-size’ enlargements

When discussing the conceptual basis for using ‘life-size’ images, Glaister and Brash [1] concluded that it was a “safer plan to make the enlargements of both the skull and portrait natural size, the former exactly so, the latter as near as might be feasible” because it would form “a more crucial experiment” than to fit the skull and face images without regard to their actual sizes. The concept underlying the use of ‘life-size’ images is that a comparison achieved by overlaying two physically

Conceptual departure when using ‘life-size’ enlargements

Ubelaker et al. [32] adjusted the size of the face-image until it filled at least 67% of the monitor screen, which departed from the use of ‘life-size’ images during superimposition. Subsequently, Delfino et al. [33] centered the skull-image on the monitor at approximately two-thirds the size of the screen. Seta and Yoshino [21] and Yoshino et al. [23] used half-size skull-images for superimposing because the skull image was focused on ground glass of that size. The work by Austin-Smith and

Conceptual basis for orienting the skull: use of points at two different planes

Glaister and Brash obtained the ‘primary’ orientation of the skull by either applying the angle of orientation of the tiara observed in the ‘life-size’ photograph of the skull or setting the skull to correspond to the salient features observed in the face in the ‘life-size’ transparencies [1]. These authors [1] recognized the requirement for slight movements of the skull to obtain its final orientation. The practice of achieving an appropriate orientation by overlaying the transparencies of the

On the use of Whitnall's tubercles during superimposition

Practitioners have long used orbital measurements from the skull to orient it relative to the posture of the face [11], [15], [24], [26], [27], [30]. However, there has been a lack of uniformity in the point in the orbital zone that is used. Morphological placement of the eye, i.e., placing it in conformity with the orbit, may shift the eye angle in the face away from the location of the Whitnall's tubercle in the orbit, as shown by some researchers [40], [42]. Chandra Sekharan [5] assumed the

Transitions in the criteria for assessing goodness of fit

Glaister and Brash relied upon salient features, such as chin height, for evaluating fit while superimposing skull tracings and face-images at ‘life-size’. Their reliance of cranial landmarks is limited – such as the use of prosthion seen exposed in the face photographs [1]. Of the seven authors reporting on photographic superimposition, six have relied on morphological features, of whom two have included asymmetries (Table 1). Of the 21 studies and case reports of video superimpositions, all

Conceptual basis for the reliance on cranial and facial landmarks

The research of Gordon and Steyn [42] has indicated, for the first time, that when other parameters are equal, the landmark method performs less reliably compared with the morphological method. Thus, when using any anthropometric landmark, the following must be considered: (a) the size of the marking of the landmarks in relation to the magnification of the images, (b) the ability to mark the landmarks on the face- and skull-images repeatedly, (c) the discriminatory power of the landmarks over

Fade and wipe resources for assessing the goodness of fit during superimposition

With regard to the fade (mix) and wipe resources available when using image-mixing devices, it has been well accepted that a very slow fade may provide the illusion of a perfect match during the period that one image is almost imperceptibly replaced by another [79], [80]. İşcan [80] cautioned that the fade approach must be avoided and recommended an initial complete wipe to be followed by a dissolving or mix analysis. Many authors have found the wipe (sweep) technique more useful [13], [15],

Reliance on asymmetries while assessing the goodness of fit during superimpositions

Arguably, in any superimposition that considers morphology, an asymmetry would naturally be included to help assess whether a match exists. However, Glassman [81] conceded that professionals who are typically trained to assess the size and conformity of anatomical structures are likely to miss aspects of contour or asymmetries that require an artistic ‘eye’ and suggests involving a facial reconstruction artist when assessing asymmetries while evaluating video superimpositions. Indeed, some of

On the use of 3D digital images of skull

Gordon and Steyn indicated that the general issue in obtaining high-quality digital scans and photographs may also contribute to the relatively poor accuracy reported in their study and considered that traditional, more laborious methods using video cameras, for example, may yield better results [42]. The authors observed missing areas in the 3D skull-images, which were attributed to the inability of the scanner to scan curved surfaces in the skull [42]. The use of digitized 3D images of skulls

Conclusion

The concepts relating to superimposition method proposed by early reports summarized here and have been accepted by several practitioners when identifying skulls in real-life cases include the use of ‘life-size’ images as a “safer” procedure [1], the use of auditory meatus-ear relationships in addition to orbital relationships while orienting the skull [5], a reliance on wipe images permitting the “better comparison of bony points” compared with the “more pleasing” mix images [13], and the

Acknowledgements

Universiti Sains Malaysia is acknowledged and thanked for the financial assistance through RU grant 1001/PPSK/813011 and Short Term Grant 304/PPSK/61312009 which supported this review.

References (95)

  • G.M. Gordon et al.

    An investigation into the accuracy and reliability of skull-photo superimposition in a South African sample

    Forensic Sci. Int.

    (2012)
  • S. Seta

    Current research and case work activities of criminalistics in Japan

    Forensic Sci. Int.

    (1996)
  • A.K. Ghosh et al.

    An economised craniofacial identification system

    Forensic Sci. Int.

    (2001)
  • R. Moreton et al.

    Investigation into the use of photoanthropometry in facial image comparison

    Forensic Sci. Int.

    (2011)
  • T. Turvey et al.

    Orbital Anatomy for the Surgeon

    Oral Maxillofac Surg Clin

    (2012)
  • J. Shi et al.

    How effective are landmarks and their geometry for face recognition?

    Computerv. Image Underst.

    (2006)
  • T. Catterick

    Facial measurements as an aid to recognition

    Forensic Sci. Int.

    (1992)
  • M. Yoshino et al.

    Computer-assisted facial image identification system using a 3-D physiognomic range finder

    Forensic Sci. Int.

    (2000)
  • J.W. Chapman et al.

    Environmental versus genetic influences on fluctuating asymmetry in the house fly, Musca domestica

    Biol. J. Linn. Soc.

    (2000)
  • S. Benazzi et al.

    Virtual anthropology and forensic arts: the facial reconstruction of Ferrante Gonzaga

    J. Archaeol. Sci.

    (2010)
  • J. Glaister et al.

    Medico-legal Aspects of the Ruxton Case

    (1937)
  • N.K. Sen

    Identification by superimposed photographs

    Int. Crim. Pol. Rev.

    (1962)
  • C. Basauri

    A body identified by forensic odontology and superimposed photographs

    Int. Crimin. Pol. Rev.

    (1967)
  • P. Chandra Sekharan

    A revised superimposition technique for identification of the individual from the skull and photograph

    J. Crim. Law Crimin. Pol. Sci.

    (1971)
  • P.C. Chandra Sekharan

    A scientific method for positioning of the skull for photography in superimposition studies

    Pol. Sci. Admin.

    (1973)
  • S. Sivaram et al.

    Identity from skeleton – a case study

    Int. Crimin. Pol. Rev.

    (1977)
  • R.B.J. Dorion

    Photographic superimposition

    J. Forensic Sci.

    (1983)
  • J.J.I. McKenna et al.

    A method of matching skulls with photographic potraits using landmarks and measurements of the dentition

    J. Forensic Sci.

    (1984)
  • R. Helmer et al.

    Vereinfachite Schadelidentifizierung nach dem superprojektionsver fahren mit Hilfe einer Video Anlage

    Z. Rechtsmedizin

    (1977)
  • H. Hagemeier

    Identification of a skull by electronic superimposition of images

    Kriminalistik

    (1979)
  • T.D. Koelmeyer

    Video camera superimposition and facial reconstruction as an aid to identification

    Am. J. Forensic Med. Pathol.

    (1982)
  • R. Bastiaan et al.

    Video superimposition of skulls and photographic portraits – a new aid to identification

    J. Forensic Sci.

    (1986)
  • P.X. Iten

    Identification of skulls by video superimposition

    J. Forensic Sci.

    (1987)
  • J.J.I. McKenna

    A method of orientation of skull and camera for use in forensic photographic investigation

    J. Forensic Sci.

    (1988)
  • O. Gruner

    Die rechtsmedizinische Identifizierung – Bedeutung und Beweiswert

    Z Rechtsmed.

    (1989)
  • P. Chandra Sekharan

    The problems of positioning skulls for video superimposition technique

    Can. Soc. Forensic Sci. J.

    (1989)
  • P. Chandra Sekharan

    Positioning the skull for superimposition

  • O. Gruner

    Identification of skulls: a historical review and practical applications

  • S. Seta et al.

    Combined apparatus for photographic and video superimposition

  • D. Austin-Smith et al.

    The reliability of skull/photograph superimposition in individual identification

    J. Forensic Sci.

    (1994)
  • A.W. Shahrom et al.

    Techniques in facial identification: computer-aided facial reconstruction using a laser scanner and video superimposition

    Int. J. Legal Med.

    (1996)
  • D. Austin

    Video superimposition at the C.A. Pound Laboratory 1987 to 1992

    J. Forensic Sci.

    (1999)
  • T.W. Fenton et al.

    Skull-photo superimposition and border deaths: identification through exclusion and the failure to exclude

    J. Forensic Sci.

    (2008)
  • B.A. Nickerson et al.

    A methodology for near-optimal computational superimposition of two dimensional digital facial photographs and three-dimensional cranial surface meshes

    J. Forensic Sci.

    (1991)
  • D.H. Ubelaker et al.

    Computer-assisted photographic superimposition

    J. Forensic Sci.

    (1992)
  • P.V. Delfino et al.

    Shape analytical morphometry in computer-aided skull identification via video superimposition

  • I. Bajnoczky et al.

    A new approach to computer-aided comparison of skull and photograph

    Int. J. Legal Med.

    (1995)
  • Cited by (0)

    View full text