Home About Us Laboratory Services Forensic Science Communications Back Issues October 2000 Abstracts from the 9th Intntl. Craniofacial ID...

Abstracts from the 9th Intntl. Craniofacial ID Conference--Part 3 (Forensic Science Communications, October 2000)

    U.S. Department of Justice
    Federal Bureau of Investigation
 FBI Seal graphic

Layout graphic: Topline




Graphic link to the Current Issue's table of contents

Graphic link to links to Back Issues

Graphic link to a search engine for all issues

Graphic link to the Journal Description page

Graphic link to the Editorial Board for this issue

Graphic link to the Instructions for Authors

Graphic link to the Handbook of Forensic Services

Graphic link to the FBI Lab Home Page

Graphic link to FBI Publications page (publications listing)

Sidebar graphic

Masthead - Forensic Science Communications
October 2000   Volume 2   Number 4

Presentations at the
9th Biennial Scientific Meeting of the
International Association for
Craniofacial Identification
Washington, DC
July 24–28, 2000

Part 3

The following abstracts of the presentations are ordered alphabetically by authors' last names.

Quality Standards for Forensic Opinions on the
Identity of Living Offenders In Pictures

F. W. Rösing
Institut für Humangenetik und Anthropologie Universitatsklinikum
Ulm, Germany

The forensic identification of persons in photographic documents follows established scientific criteria and methods. The recognition of human faces, however, is a highly developed, natural human skill. The widespread belief that a professional forensic identification is as easily accomplished as common facial recognition among individuals is problematic. In response to the problems generated by this mistaken belief, a group of university-based experts has formulated a set of quality criteria for such identification opinions. The goal of these criteria is the creation of an identification system that can be peer-reviewed for expertise. The criteria are as follows:

  • The main principle of photographic identification is the assessment of detailed structures in a slow and thorough process and with a gradation of the result, as opposed to holistic, fast, and polarizing daily recognition.

  • The country-specific judicial basis must be considered.

  • Photographic cameras are recommended for surveillance equipment as video cameras have too low a resolution. A surveillance camera should not be installed too high above the scene because this reduces recognition of the lower face. For an identification, comparison pictures of the suspect should be made, preferably with the original security camera. An identification should be performed by picture-to-picture comparison and not by a picture-to-person comparison.

  • The picture quality influences recognition. This is an independent variable that must be assessed in every extracted trait parallel to the assessment of the trait itself.

  • Identifying traits are morphological structures that can be seen on the surveillance picture, particularly those of the face and ears, but also of the thorax or hands, or of stature and posture. Population frequencies are decisive in the assessment of traits, and the possibility of a secondary change of traits (e.g., through use of a mask or other means of obfuscation) should be considered.

  • The expert opinion may explain the scientific bases of photographic identification. All traits found must be described, preferably using established anthropological nomenclature, but with due regard that the opinion should be understood by non-experts. Partial expert commissions—for example, of an ear-should not be accepted. All principles used and assumptions made must be described through every step of the identification process. The result of an expert opinion is the identity probability.

  • The identity probability depends upon the number and rarity of the traits found. If traits are independent from each other, the single probabilities are multiplied according to the usual probabilistic rules; if the traits are correlated, the covariance reduces probability. If trait frequencies are not formally known, they should be estimated. For the final probability result, a verbal formulation should be given, possibly also a figure, together with the range.

  • The principle of preselection must be considered. The normal probability assessment depends upon the assumption that a suspect is a chance draw from the general population. If he or she has been found using the surveillance photograph, this is not a chance draw. Then every suspect named is similar to the offender. Consequently, the population-based probability assessment can not be applied, but the demands for rarity of traits increases, and traits that are normally not included in the recognition of persons, such as the ear traits, become important.

  • As a general precaution, the possibility that a close genetic relative might also be suspected should be checked. He or she should be included in the identification process.

  • The scientific basis of photographic identification is physical anthropology, which includes as a subfield the morphology of living persons. The assessment of external traits formerly used to diagnose paternity is particularly useful as a basis. A medical doctor or criminalist might acquire and develop the necessary knowledge and experience of physical anthropology during work on scene or in the field.

Several procedural steps are recommended for quality assurance, including regular determination of the intra- and inter-observer error, the exchange of model cases, consultation with colleagues before opinions are issued, and the examination of new members.


Buhmann, D., Helmer, R. P., Jaeger, U., Jürgens, H. W., Knussmann, R., Rösing, F. W., Schmidt, H.D., Szilvássy, J., and Ziegelmayer, G. Standards für die anthropologische Identifikation lebeder Personen auf Grund von Bilddokumenten. Grundlagen, Kriterien und Verfahrensregeln für Gutachten, Anthrop Anz (1999) 57(2):185–191. (Also published in Dt. Autorecht [April 1999], pp. 188–189; Kriminalistik [April 1999], pp. 246–248; Neue Z. Strafrecht NStZ [1999] 5:230–232; Rechtsmed [1999] 9:152–154.)

Back to the index

Radiographic Identification: Importance and Cases

F. W. Rösing, R. Novak, T. Fleiter, R. Hausmann, and K. W. Alt
Institut für Humangenetik und Anthropologie Universitatsklinikum
Ulm, Germany

The differential value of the dentition in the identification of unknown skeletons has been reevaluated in several recent studies. These studies included a thorough check of existing dental records in which more than one third of all single file entries were either wrong or missing (Hausmann et al. 1997). Moreover, the publication of the dental records of unknown recovered bodies yielded an identification success rate of only five percent (Alt and Walz 1999). These data challenge the widespread opinion that teeth have a decisive value for human identifications and add new weight to the ancient practitioner's rule of using every technique available for identification.

In northern countries, the classic technique of radiology is used as an alternate means of identification. For various diagnostic purposes a large portion of the population has been X-rayed, and the films are normally stored. Skull, thorax, spinal column, and pelvic girdle are the skeletal regions most frequently used in comparisons, but the clavicle, sternum, patella, and calcaneus are diagnostic as well (see Case A, below). The strong individuality of the anatomical structures depicted in these radiographs and the clear and direct diagnostics of this technique are highly valued for identification purposes.

Problems associated with radiographic comparisons include:

  • Low quality clinical films, such as those revealing little antemortem evidence;

  • Variability in the orientation of skeletal remains between the clinical and postmortem images;

  • The presence of overlapping trabecular structures (Case B);

  • The presence on ontogenetic changes (Case C); and

  • The digital storage of Roentgen films, which allows the possibility of image manipulation and a consequently drastic reduction in their value as forensic documents.
Case A: The Identifying Vertebra

A mummified body was found hanging in a tree 11 years after death. A missing man was linked to the body following analysis of age, sex, and stature, but no dental records were available and impeding mummified tissues made facial identification imprecise. Final identification was achieved through analysis of radiographs showing identifying traits in the lower vertebrae and sacroiliac fusion.

Case B: The Shredded Body

Parts of a fresh body were found in a garbage processing plant
between the coarse and fine shredders. Sexing and aging of the remains were difficult, and there were no remaining teeth. Portions of the jaw were present, however, and identification was possible using the trabeculae of the mandibular body.

Case C: Blurring Growth

The skeletonized body of a young person was found in a forest approximately 18 months after death. The sex, age, and stature of the skeleton were linked to that of a missing girl for whom no dental records were available. Identification, complicated by ontogenetic changes visible in the elbow region, was accomplished through analysis of strongly individual facial traits.


Hausmann, R., Liebler, M., and Schellmann, B. Dental records as an aid to identification: Quantity and quality of dental documentation, Rechtsmedizin (1997) 7(3):88.

Alt, K. W. and Walz, M. Dental print media and their value if forensic odontology, Journal of Forensic Odonto-Stomatology (1999) 17(1):5–9.


Preliminary Study on Facial Soft Tissue Thicknesses by
Magnetic Resonance Imaging (MRI) on Northwest Indians

D. Sahni, I. Jit, D. Sanjeev, P. Singh, S. Suri, and H. Kaur
Postgraduate Institute of Medical Education and Research
Chandigarh, India

It is not uncommon for the police to need to identify the deceased from a skull. If data on facial soft tissue thicknesses is available, a facial reconstruction may be done. Work of this nature has been performed in Western and Japanese populations (Rhine and Campbell 1980), but the methods used lacked adequate reliability, and the data generated from one population cannot be extrapolated to another population. No work on this subject has been done in India, where previously only the approximate age and sex of the skull (Jit 1979) and sometimes the cause of death (Jit 1993) could be determined. Lack of data for the Indian population resulted in the inability to accurately reconstruct facial features from skeletal remains, and, consequently, many murder cases could not be traced. It is important to find out the facial soft tissue thicknesses in various populations in India.

In Japan and Western countries, the following methods have been used for measuring facial soft tissue thicknesses:

  • Needle depth probe method on cadavers (His 1895; Kollman and Buchly 1898);

  • Radiography, including computerized tomography (CT) scan (Aulsebrook et al. 1996; George 1987; Philips and Smuts 1996);

  • Ultrasonography (Lebedinskaya et al. 1993); and

  • Magnetic resonance imaging (MRI).

MRI offers several advantages. It allows multiple planes for study, thus allowing accurate measurements perpendicular to the area of interest. A greater number of anatomical points can be studied accurately using MRI as compared to CT scans. The soft tissue details are much better on MRI as compared to CT scans. In the present preliminary study, measurements of soft tissue thicknesses by MRI were taken at 29 fixed anatomical points in 30 male and 30 female northwest Indian adults. The following sections were taken into consideration to measure the tissue thicknesses at various anatomical points in the MRI:

  • Sagittal section studies included supraglabella, glabella, nasion, end of nasal (rhinion), midphiltrum, upper lip margin, lower lip margin, chin lip fold, mental eminence, beneath chin, vertex, inion, and opisthocranion.

  • Parasagittal section at the level of mid-orbit studies included frontal eminence, supraorbital, and infraorbital on both sides.

  • Coronal section studies included right porion, left porion, right gonion, and left gonion.

  • Transverse section studies included right zygion, left zygion, right outer canthus, left outer canthus, right supra M2, and left supra M2.

  • The additional eight landmarks measured were not studied by other authors.

Mean values of the measurements show that in the northwest Indian population, the measurements are quite different from that described in European populations (Rhine and Campbell 1980). Differences were particularly noted in the landmarks at glabella, nasion, midphiltrum, mental eminence, supraorbital, infraorbital, and zygion.


Aulsebrook, W. A., Becker, P. J., and Iscan, M. Y. Facial soft tissue thicknesses in the adult male Zulu, Forensic Science International (1996) 79:83–102.

George, R. M. The lateral craniographic method of facial reconstruction, Journal of Forensic Sciences (1987) 32:1305–1330.

His, W. Johann Sebastian Bach's Gebeine und Anlitz nebst Bemerkungen uber Dessen Bilder, Abhandlung durch Mathematik und Physik (1895) 22:380–420.

Jit, I. The House of Skeletons case, Journal of the Anatomical Society of India (1979) 28:106–116.

Jit, I. Case report: Dr. Mela Ram's murder case in Chandigarh, Journal of the Indian Academy of Forensic Medicine (1993) 14:11–14.

Kollmann, J. and Buchly, W. Die Persistenz der Rassen und die Reconstruction der Physiognomie Praehistorischer Schadel, Archiv fur Anthropologie (1898) 25:329–359.

Lebedinskaya, G. V., Balueva, T. S., and Vaselovskaya, V. S. Principles of facial reconstruction. In: Forensic Analysis of the Skull: Cranio Facial Analysis, Reconstruction and Identification. Eds. M. Y. Iscan and R. P. Helmer. Wiley-Liss, New York, 1993, pp.183–198.

Phillips, V. M. and Smuts, N. A. Facial reconstruction: Utilization of computerized tomography to measure facial tissue thicknesses in a mixed racial population, Forensic Science International (1996) 83:51–59.

Rhine, J. S. and Campbell, H. R. Thickness of facial tissues in the American Blacks, Journal of Forensic Sciences (1980) 25:847–858.

Back to the index

Variation of the Glabella Profile by
Roentogenographic Cephalometry in Northeast Asians

R. Sakashita, N. Inoue, and T. Kamegai
Kumamoto University
Kumamoto, Japan

There are many variations of the glabella profile used to identify persons and racial origin. Among the Japanese, two major facial types are hypothesized: the Jomon type with prominent glabella and the Yayoi type with flat glabella (Hanihara 1999). This study was conducted to test this hypothesis and also to discover the range of variation in northeast Asians.

Materials and Methods

A total of 812 male crania, comprised of the following, were used:

  • 18 Japanese groups from the northeastern to the southern islands of Japan, from the Jomon period (8000–3000 YBP) to the present, housed at the University of Tokyo, Kyoto University, Kyushu University, and Nagasaki University;

  • 1 Korean group housed at Kyoto University; and

  • 2 Chinese groups from the northeast to the root of the Yellow River, from 7000 YBP to 1800 YBP, housed at the Chinese Academy of Social Science.

X-rays of all possible cases were used. To clarify the shape of the glabella, seven angular measurements and seven linear measurements (Inoue et al. 1997) were made. The significant difference among group means was tested using GLM and Tukey's procedures. Fast-cluster was used to classify the samples into two groups: the prominent glabella group and the flat glabella group.

Results and Discussion

The means of all measurements among groups were found to be significantly different by GLM (p < 0.0001). The distance from the nasion to the tangent line of nasal bone at the nasion was greater among Japanese groups (M = 5.5–11.9 mm) than among Chinese and Korean groups (2.2– 5.2 mm), with the exception of Japanese groups from the Yayoi and Kofun periods (2200–1500 YBP, 3.1–5.9 mm). The distance from the nasion to glabella-metopion was greater among the eastern Japanese groups in the Jomon period and the upper Yellow River (8.7–9.7 mm) than among groups from the Kofim period (4.8–7.7 mm), other groups of the Yellow River, and Korea (6.8–7.7 mm). The tangent of frontal bone at the nasion to glabella-metopion was acute among the eastern Jomonese and wide among groups in the Yayoi and Kofun period.

In the Jomon period, most samples from the northeast and east were classified into the prominent glabella group; however, the rate was reduced in the west. In the Yayoi period, more than 70 percent of samples in Kyushu were classified into the flat group; in the southern island, 63.6 percent were classified into the prominent glabella group. In the Kofun period, most Japanese samples were classified into the flat groups. In the Medieval period, almost 60 percent of samples were classified into the prominent glabella group. More than 70 percent of the Chinese groups were classified into the flat groups; however, in the upper Yellow River, the rate was reduced to 30 percent. All Korean samples (n = 10) were classified into the flat group.

Compared to Chinese groups, in which the majority of samples were classified into the flat group with few exceptions, there were variations among Japanese samples. The shape of the glabella would be affected by the pressure of the frontal sinus, which might be affected by the environment. It was suggested from this result that the shape would not be related to the environmental temperature or moisture. The variation in Japanese glabella shape fits the hypothesis of a migration of people with a paddy-rice agriculture from Korea to eastern Japan at the time of the Yayoi period. However, it is uncertain why there were the flat glabella groups in Kyushu in the Jomon period.


Hanihara, K. Jomonese face and Yayoi face (in Japanese). In: Scientific Approach to the Face. Eds. Baba, H. and Kanawawa, E. Newton, Tokyo, 1999, pp. 68–85.

Inoue, N. Materials and methods of investigation. In: Tooth and Facial Morphology of Ancient Chinese Skulls. Eds. Inoue, N. et al. Therapeia, Tokyo, 1997, pp. 17–34.

Back to the index

Identification of Sex From the Quantification of the Mandibular Size and Shape Through Use of
New Developments in the Elliptic Fourier Method

M. Schmittbuhl and P. Mangin
Institut Universitaire de Medecine Legale
Lausanne, Switzerland

M. Perrier
Policlinique Medicale Universitaire
Lausanne, Switzerland

Sex identification has been investigated in nearly every part of the skeleton. However, data dealing with sexual dimorphism of the mandible remains relatively rare in the literature. The few existing studies attempted to study the sexual dimorphism of the mandible using classical metric measurements and revealed a sex identification accuracy ranging from 75 percent to 84 percent. At the present time, none of them takes into account the complete shape of the mandible with regard to the differentiation of sex.

The aim of the present study was to examine the sexual dimorphism of the mandible from the quantification of its shape using new developments in the elliptic Fourier method. The sample consisted of a series of 138 lateral views of toothed mandibles (77 males, 61 females) from individuals of known sex aged 20 and older. The outline of the mandible was automatically extracted by image analysis procedures and characterized by an ordered series of harmonics. Each harmonic was described by four new parameters termed elliptic descriptors. A characterization of the shape of the mandible was possible with great precision using elliptical Fourier analysis. The study of the shape alone indicated that the average accuracy of sex determination was 86.44 percent. The study of both size and shape revealed that the average accuracy of sex determination was 95.24 percent.

The superiority of the present results compared to those obtained by previous studies can be explained by the fact that conventional morphometric methods based on linear measurements provide only partial shape information of the mandibular outline. From the study of the morphological contributions of the elliptical descriptors, the present analysis revealed that sex differences could be explained by both size and shape differences in the mandible, with differences in shape focusing principally on the ramus, the condyle, and the robustness of the mandibular body. The present method, which relies on elliptical descriptors, allows a precise characterization of the sexual dimorphism of the mandible for the first time and should constitute an interesting approach for an accurate sexing of fragmented skeletal material.

Computer Techniques for Comparing Antemortem and Postmortem Dental Radiographs Using
Digitized and Digital Images

D. R. Senn, P. Brumit, M. White, and M. E. Alder
Center for Education and Research in Forensics
University of Texas Health Science Center
San Antonio, Texas

Dentists are taught from the beginning of their training to recognize anatomical, pathological, and restorative features in dental radiographs. This skill, when cultivated over years, helps forensic dentists to make comparisons of features in dental radiographs to assist in determining identities. Computer techniques allow direct visual comparison by overlaying one image on another, decreasing the necessity for subjective judgments and enhancing scientific objectivity. These techniques also facilitate demonstration to nondentists in an "easy-to-see" manner and may be useful in trial settings to demonstrate identifications to triers of fact. This presentation demonstrates two objectively reproducible computer techniques that facilitate direct comparisons between antemortem and postmortem radiographs.

Antemortem radiographs are currently most often received as films. With the rapid growth of the use of digital radiography in dental offices, the day may come when most antemortem radiographic information will be digital. Comparisons in which the only antemortem radiographic information available was in the form of digitized images received by E-mail have already been made. The advantages of digital versus film radiography have already been documented in publications, and in forensic examinations these advantages are substantial.

Digital radiographs were acquired using a Schick CDR digital dental unit with a size 2 sensor. Digitized film images were acquired from standard dental radiographic films using an Epson Expression 636 flatbed scanner with transparency scanner. Digital or digitized film antemortem and postmortem radiographs were opened in Adobe Photoshop 5.0 using a Dell Dimension XPS R400 computer with a 400-megahertz Pentium II processor. Using the Twain acquire feature in Adobe Photoshop and adjusting the scanning resolution to match that of the digital images, films were imported. Using landmarks in restorations or anatomical features, the measure tool, and commands from the toolbar, it was possible to assure that the features to be compared in both images were compatible in size.

Using the Image>Adjust commands and the Magic Wand tool, individual restorations in the postmortem image were depicted as either solid volume or hollow volume images. These images were transferred to the antemortem image. Using the Edit>Transform>Rotate commands and the Move tool, the postmortem image was aligned in the same orientation as the antemortem and the direct comparison accomplished.

The transparency overlay comparison technique allows the overlay of transparent images of postmortem radiographs over nontransparent antemortem images. The transparency factor can be adjusted to allow a fade-in, fade-out comparison, and the transparent image can be moved onto or away from the underlying antemortem image. This presentation demonstrates both methods from start to finish and handouts supply step-by-step instructions.

As in earlier nondigital methods, exposing the postmortem radiographs in the same or similar orientations used in the antemortem is essential. Cases with metal restorations are best suited to the hollow volume technique. The transparency overlay method works well in cases with or without restorations. Both methods seem complicated when described but become intuitive when practiced. Postmortem radiographs made on "John Doe" cases may not be suitable for these techniques if subsequently acquired antemortem radiographs were not exposed in the same orientations.

Direct computer comparison of antemortem and postmortem digital or digitized radiographs allows for direct and objective comparison instead of indirect or side-by-side subjective comparison. These techniques allow the investigator to illustrate more scientifically reproducible comparisons. The techniques are used as teaching techniques for postdoctoral forensic dentistry students and facilitate dental identification demonstrations to medical examiners and coroners. They may be useful tools in the courtroom to illustrate identifications to judges and juries or help to verify identity in disputed identity cases.

Back to the index

Significance of Hereditary Factors in
Forensic Ear Identification

S. A. Sholl and P. Vanezis
University of Glasgow
Glasgow, Scotland

There has been relatively little speculation as to the potential of ears as a means of identification, and this possibly is due to the fact that there are so many other ways to identify an individual (Pertson 1987; Moenssens et al. 1995; Tilley and Ford 1996). Several attempts, notably by Alphonse Bertillon in 1862 and Alfred Iannarelli in the 1940s (Iannarelli 1964 and 1989), have been made to establish a classification system for the ear with varying success. Generally speaking, criteria are prioritized by sex and race, followed by measurements of the dimensions of the ear and the relationship between various anatomical and anthropometric landmarks.

A study was undertaken to ascertain whether there is any significant association between ear characteristics in same-sex or opposite-sex members of the same family. Two-dimensional digital images of left and right ears are used to compare the configurations of four preselected anatomical points: the intertragic notch, the tragus, the antitragus, and the malchus. The malchus refers to the point on a 2D image of the ear where the curve of the antihelix intersects with the curve of the helix root. It has been named "malchus" after the servant mentioned in the Gospel of John whose ear Peter cut off (John 18:10). A method was devised to standardize the digital images to the extent that any differences in the angles at which they are taken have no influence on the configuration being studied. Anatomical points are identified and analyzed using software designed by the University of Glasgow.

Preliminary analysis using this ear configuration indicates a physiological relationship between female members of the same family, but it indicates no relationship between male members of the same family or between male and female members.


Iannarelli, A. V. Iannarelli System of Ear Identification. Foundation, New York, 1964.

Iannarelli, A. V. Forensic Identification Series: Ear Identification. Paramont, California, 1989, p. 5.

Moenssens, A. A., Starrs, J. E., Henderson, C. E., and Inbau, F. E. Scientific Evidence in Criminal Cases (4th ed.). Foundation, New York, 1995, Chapter 9.

Pertson, W. J. Ear, lip and nose prints: Another means of identification, RCMP Gazette (1987) 49(9):7–15.

Tilley, N. and Ford, A. Forensic science and crime investigation. In: Crime Detection and Prevention Series Paper 73. Home Office, London, 1996.

Do Resemblance Ratings Measure a
Facial Approximation's Accuracy?

C. N. Stephan
Department of Anatomical Sciences
University of Adelaide
Adelaide, Australia

Facial approximation is a method used to build a deceased person's face from his or her skull to indicate their facial appearance before death. In forensic science, facial approximation is used to promote recognition of a deceased person in an attempt to generate leads that may aid the process of identifying skeletal remains.

Because forensic facial approximation is used to promote recognition, an accurate forensic facial approximation should be easily recognized as the person to whom the skull belonged—the target individual. However, attempts to assess the accuracy of forensic facial approximations have commonly been made by resemblance ratings of the similarity between a facial approximation and the corresponding target individual.

This study tests the validity of using such resemblance ratings to assess the accuracy of forensic facial approximation. Results of the study indicate that there is no statistically significant difference between resemblance ratings of facial approximations to target individuals and resemblance ratings of facial approximations to individuals incorrectly identified as target individuals.

It is concluded that a resemblance rating of a facial approximation to an individual does not indicate the facial approximation's accuracy because a nontarget individual may receive a resemblance rating equal to or greater than the target individual.

Digital 3D Reconstruction of Skulls From
Fragments Using Streifenlichttopometrie (SLT) and a Special DNA Method

J. Subke, M. Zeller, H. D. Wehner, and H. Wolf
Institute of Legal Medicine
University of Tuebingen
Tuebingen, Germany

This project aims to develop a method of 3D reconstruction of skulls from fragmented skeletal remains using Streifenlichttopometrie (SLT) as a digital high-resolution 3D documentation method and a special DNA method applicable for use on bones up to Neolithic age for fragment differentiation.

In cases involving more than one victim, the fragmented skeletal remains are first differentiated by the Zeller method for the extraction of DNA from skeletal material up to Neolithic age. Isolated and concentrated DNA can be submitted to the PCR of nuclear STR-loci and of the mitochondrial D-loop sequence. Through STRs and mtDNA analysis, the skeletal remains are classified and a possible kinship can be characterized.

In the second step of the reconstruction, the classified fragments of the skull are digitally documented in 3D by Streifenlichttopometrie (SLT), a method introduced by Subke in 1998 as a new 3D photorealistic color documentation for forensic medicine. SLT is an adaptable, optical 3D measurement system, which allows the fast (1000,000 point/sec.), high-resolution(< 1 mm) measurement of surfaces and includes the natural color.

In the third step the skull is digitally restored from the fragments documented in 3D. With the help of an efficient algorithm of the visualization software, the exact matching of neighboring fragments is possible. A reference skull with anatomical landmarks is used to support the exact positioning of the fragments. The reconstruction of missing fragments is achieved by means of surface interpolation, the symmetrical properties of the skull, and a skull database in order to complete the reconstruction.

The application of SLT provides the following advantages to digital 3D reconstructions:

  • Fast 3D measurement,

  • Submillimeter resolution, and

  • Storing of geometry and color.

In combination with the Zeller method of DNA extraction, which allows the differentiation of bones up to Neolithic age, the SLT method is not only effective for the identification of recent cases, but also for cases of anthropological and archaeological interest.

Back to the index

Effect of Various Facial Heights in an
Edentulous Skull With Special Reference to
Police Artist Reproduction

M. Taister and J. A. DiZinno
Federal Bureau of Investigation
Washington, DC

H. Borrman and G. E. Carlsson
Göteborg University
Göteborg, Sweden

When a skull from a deceased, unknown individual is found by the police, the teeth are usually present in the jaws, which facilitates identification. For a skull of an edentulous subject, facial reproduction is much more difficult because it is not possible to determine the facial height. This issue has received little attention in the forensic literature. Facial height is closely related to the vertical dimension of occlusion (the height of the bite when the teeth are biting together). The aim of this experiment is to illustrate the effect of variation in facial height on a police artist's facial reproduction of an edentulous skull.

The true facial height values for any edentulous person are unknown. For an edentulous skull, three sets of occlusal rims (devices used for the determination of vertical and horizontal jaw relations in the fabrication of complete dentures) with different vertical dimensions were constructed in a dental laboratory. These rims produced three different facial heights on which the police artist based the facial reproduction. One set provided a facial height that was determined according to average dental and facial measurements and related to the skull dimensions of this case. The other two sets of rims were designed to establish one lower and one higher facial height than the first with an assumed average dimension.

The three reproductions will be demonstrated and discussed with regard to the effect of varying vertical dimension of occlusion on the police artist's facial reconstruction of an edentulous person.

Gross Morphological and Visual Examination Versus
DNA Technology: Who Do You Trust?

K. T. Taylor
Facial Images
Austin, Texas

D. M. Glassman
Southwest Texas State University
San Marcos, Texas

This case study examines the issue of placing total trust in an established scientific methodology even though gross morphological and visual assessment indicate contrary findings.

In August 1997, forensic artist Karen Taylor was asked by the San Antonio, Texas, Police Department to prepare 2D facial reconstruction drawings from a skull found in the backyard of a San Antonio residence. Taylor was told that the wife of the man who lived at the residence had been missing for several months but had been ruled out as a match to the skull by DNA analysis.

In addition to the skull, mandible, and fragmentary postcranial elements, a clump of human hair was recovered at the scene. Preliminary examination of the compressed mass indicated that it was composed of diverse hairs of various lengths, colors, and textures. Lacking reliable hair information, the reconstructions were prepared with an intentionally generic hairstyle and color to allow for viewer interpretation. Taylor developed both frontal and lateral projections of approximate facial appearance in life with some input from skeletal biologist Dr. Gentry Steele of Texas A&M University, who was involved in the recovery at the scene. The drawings were distributed to the news media in late August, but no responses resulted in the identification of the remains.

In September 1997, forensic anthropologist David M. Glassman was asked to examine the same skeletal material found at the San Antonio residence. The condition of the skull was excellent and complete except for a dissected region of the occipital and two teeth, which were extracted for DNA analyses. The postcranial material was comprised of several hundred highly fragmentary pieces. The fragmentation was determined to be the result of perimortem sharp and crushing trauma over all areas from the neck to the feet.

Osteologic analysis of the cranium and mandible revealed a female adult of European ancestry. The postcranial fragments were sorted by bone element wherever possible. All identifiable fragments were compared for duplication to assess the MNI statistic. No instances of duplication were observed. The fragments were relatively consistent in age and sex morphology, resulting in a conclusion that the remains represented a single individual. Portions of three cervical vertebrae including the atlas, axis, and C3 were available to match with each other and with the occipital condyles of the cranium. Consistency of articulation was found. The osteologic exam concluded that there was no indication that the remains represented more than one individual.

In November 1997, Taylor viewed a segment of the television program, Unsolved Mysteries, which profiled the San Antonio case and included a photograph of the missing wife. Taylor, who was unaware of Glassman's findings and had not previously seen photographs of the wife, was startled by the morphological resemblance of the wife to the reconstructions. She inquired whether the missing woman possibly had a sister. Upon learning that she did not, Taylor did rudimentary overlay comparisons of the skull with photographs of the missing woman. Two different facial photographs indicated a disturbingly favorable but unscientific comparison, in light of the DNA findings.

The initial DNA testing of the skull, teeth, and postcranial samples indicated the possibility of three or more individuals, with the DNA of the wife reportedly found in the fragments but not in the skull. This led to a prolonged investigation focused on discovering the identity of the additional victims. Ultimately, the skull and the fragmentary remains were retested by the Armed Forces Institute of Pathology and proven to be those of the missing wife.

Anatomical Facial Sculpting With Clay and Computers:
A Case Report—Triumph for DNA

R. G. Taylor and J. G. Clement
School of Dental Science, University of Melbourne
Melbourne, Australia

The discovery of skeletonized or decomposed human remains creates major problems for investigators attempting to identify a victim. Often there are no clues to the identity of the deceased, and despite thorough investigations by the police, the remains are not claimed. When these traditional methods fail, an anatomical 3D facial approximation of the deceased may be attempted. A thorough anthropological examination is carried out to estimate the sex, age, and race of the unknown individual using skeletal and dental features. If the facial skeleton is damaged or incomplete, a preliminary reconstruction of the bony skeleton will be required. The dentition and/or dentures (or lack of them in this case) should be closely examined, as they can provide valuable clues to the articulation of the mandible in the maxillary fossae. It is also possible to note loss of vertical height from the face caused by excessive wear of the occlusal surfaces. Bone loss from the maxillary and mandibular ridges, abnormal posturing of the mandible, and lack of soft tissue support of the lips may occur, especially in the wearers of old dentures. Some of the most prominent features of the face are reliably inferred from the skeleton when sculpting, but others are not. The tip of the nose, the shape of the eyebrows, and the size and shape of the lips and ears are difficult to deduce. Any manual alteration to these features is at best a matter of trial and error, not to mention time-consuming. Enhancement through use of a computer and its associated software can easily overcome this problem. Stored images with slight variations allow comparison without alteration to the original. The final result is a humanized image that can be easily manipulated to create alternate likenesses, such as an unshaven or bearded face.

A Disaster Victim Identification Training Kit for
Odontologists and Scene Investigators

R. G. Taylor, P. J. G. Craig, and J. G. Clement
University of Melbourne, Australia
Melbourne, Australia

J. Ceddia
Monash University, Australia
Melbourne, Australia

V. Winship
Victorian Institute of Forensic Medicine
Melbourne, Australia

As part of a program for increased preparedness to combat and mitigate the effects of mass disaster on the Australian community, the inaugural National Disaster Victim Identification Workshop was hosted in February 1997 by Emergency Management Australia at Mt. Macedon, Victoria. It was decided during the workshop that each State in Australia would be responsible for preparing resources sufficient to train personnel for any type of disaster resulting in the deaths of up to 50 people.

The forensic odontology group at the Victorian Institute of Forensic Medicine and School of Dental Science, University of Melbourne, were provided subsequent support by the National Institute of Forensic Science (NIFS) to produce a set of anatomically faithful replicas of skeletal remains for the training of odontologists and other scene personnel.

The training set was derived from six original sets of human cranial remains that were carefully selected to provide a range of ages and ethnic traits. Replicas of the original remains and modifications of them were made using a silicone rubber split mold technique. Final replicas were made in pigmented methyl (colored methacrylate) to provide a realistic appearance. Partial remains were fashioned from broken pieces of larger replicas. A variety of dental treatments were performed on natural teeth substituted in the replicas. Odontograms for all treatments were constructed and entered into a custom-built computer software package specifically designed to handle dental data in large fatality situations.

The software 'DAVID' (Disaster And Victim IDentification), which includes the replicas and odontograms, is available on CD-ROM and is also downloadable from the World Wide Web at the following URL:


Radiographs of the replicas are also supplied.

The entire package is available for hire from the NIFS by police, emergency services, universities, and institutes of pathology in all states and territories of Australia. It is intended for use by local odontologists for the training of colleagues less experienced in disaster victim identification and can be used as a large table-top exercise or in field situations where it can be combined with animal carcasses and clothing to give greater realism.

Back to the index