Abstracts from the 9th Intntl. Craniofacial ID Conference--Part 2 (Forensic Science Communications, October 2000)
Presentations at the
A large series of well-documented, naturally mummified individuals were found during reconstruction work at the Dominican Church in Vác, Hungary, between 1994 and 1995.
A facial restoration of one of the mummies (right, top and middle), Anital Simon (17721808), a well-known priest, teacher, and the director of the Institute of the Deaf in Vác, was made. An authentic portrait of Simon dating from the nineteenth century exists (right, bottom). However, the portrait was not viewed until the reconstruction was complete (see models, bottom left and middle). The portrait was then compared to the reconstruction.
The portrait provided a good assessment of the reliability of the facial reconstruction. The quality and life-likeness of a facial reconstruction can be judged by using the superimposition technique to directly compare the features of the portrait with those of the reconstructed face (as seen in the photograph, bottom right). With this technique, topographical and anatomical relationships between the skull bones and soft tissues, the proportions of the skull and face, and other morphological aspects that combine to give the face its individuality can be observed.
The facial reconstruction was compared with the portrait, and the following results were obtained: 62 percent of the characteristics showed great resemblance to the original, 35 percent close resemblance, and 3 percent approximate resemblance.
Y. Lan and Y.
Tieling Research Institute
Liaoning, People's Republic of China
Tieling Public Security Bureau
Liaoning, People's Republic of China
The TLGA-213 Skull Identification System consists of a computer, a CCD camera, a video synthesizer, a skull adjustment device, and a standard platform. An image-processing card and a multimedia card are provided with the computer. The computer can capture a digital image and pass it through to the video synthesizer. A superimposed image of the digital image and the analog image from the camera will appear on the camera.
This system provides four display options: superimposed image, vertical section of the superimposed image, horizontal section of the superimposed image, and single image (skull/photograph). When the superimposed image is displayed, the brightness of the two images can be adjusted for better observation. It also has the function of false color and partial enlargement.
During the course of identification, first adjust the camera, pointing the lens face straight down. Place the photograph under the lens on the platform filling the screen. Adjust the image scale, focal length, and aperture. When the object distance is within the range of one meter, adjust the zoom lens to acquire a real image. Transfer the dynamic image into a static one by using the software program. Draw marking lines on the photograph. At this point, the photograph and the marking lines can be saved to the disk.
Adjust the angle of the camera to aim at the skull, then select the mode of superimposition display, and a superimposed image of the photograph and the skull will appear on the monitor. Examine and measure item by item in accordance with the principles and methods for skull-image superimposition identification. Finally, the identification conclusion will be provided by the computer automatically.
Following exhumation of the body of Johann Sebastian Bach from St. John's Cemetery in Leipzig, the skull was compared to several paintings. Among these were an oil painting by Elias Gottlieb Hausmann (1747) and other portraits, including one by a relative (Gottfried Friedrich Bach, about 1735), and one by an unknown painter who gave Bach a receding forehead, which, judging from the skull, the composer had.
Using the superimposition technique, differences in the face including low eye sockets, a deeply set nose, a strong skull, a projecting lower jaw, and a ptosis of the right upper eyelid characterized the probable appearance of J. S. Bach.
Cattedra di Antropologia, Università degli Studi di Bari
T. Cipriani and
V. Pesce Delfino
Consorzio di Ricerca DIGAMMA Bari
M. Colonna and
G. Di Vella
Cattedra di Medicina Legale, Università degli Studi di Bari
Many techniques have been proposed for the identification of unknown skeletal remains by comparison between a skull and a photograph taken while a disappeared subject was still alive. To be presented in the courts, however, craniofacial superimposition requires numerical parameters to be evaluated in terms of "true" or "false" attribution of the found skeletal remains (Pesce Delfino et al. 1986). The parametrization of results to decrease subjective evaluation and obtain numerical values for systematic comparison was achieved using the Shape Analytical Morphometry (S.A.M.) Forensic workstation (Colonna et al. 1980; Pesce Delfino et al. 1993). The S.A.M. approach is based on the extraction of shape-descriptive parameters by means of analytical procedures derived from analytical geometry (Lestrel et al. 1974; Pesce Delfino et al. 1997) that are therefore given the generic term "analytical morphometry" (Pesce Delfino et al. 1983). These rigorously standardized procedures allow the observation and comparison of the profiles of corresponding cranial segments. This comparison yields an analytical description of the curve profile of the skull and photograph, respectively, and provides numerical parameters with descriptive and consequently comparative meaning.
It is important to stress that the only useful parameters are those that can describe shape because discrete measurements and their corresponding derived fractions cannot describe small local differences in features that allow identification.
The numerical results are arranged in a "final evaluator" using four S.A.M. parameters: one from original curve and fundamental curve match, one from parabolic fitting, and two from Fourier analysis. Because S.A.M. provides Fourier coefficients that provide independent amplitude and phase values of each harmonic contributor, the amplitude and phase of Fourier contributors were used as correlation coefficients matching skull and photo series (linear correlation for amplitude and Pearson correlation for phase). The lowest values were obtained in true comparisons.
The proposed quantitative procedures are of the continuous analytical type and are not represented by discrete measurements between the pairs of points. As the profile is subdivided into many points, it contains a great deal of extractable information. In conclusion, this does affect the value of the final numerical evaluators showing the highest value in false comparison and the separation of "true" and "false" comparisons of both oblique and lateral view profiles.
Colonna, M., Pesce Delfino, V., and Introna, F. Identificazione mediante sovrapposizione cranio-foto del viso a mezzo di circuito televisivo: Applicazione di una nuova metodica, Bollettino della Società Italiana di Biologia (1980) 56:22712276.
Lestrel, P. E. Some problems in the assessment of morphological size and shape differences, Yearbook of Physical Anthropology (1974) 18:140162.
Pesce Delfino, V., and Ricco R. Remarks on analytic morphometry in biology: Procedure and software illustration, Acta Stereologica (1983) 2:459464.
Pesce Delfino, V., Colonna, M., Vacca, E., Potente, F., and Introna, F. Computer-aided skull/face superimposition, American Journal of Forensic Medicine and Pathology (1986) 7(3):201212.
Pesce Delfino, V., Vacca, E., Potente, F., Lettini, T., and Colonna, M. Shape analytical morphometry in computer-aided skull identification via video superimposition. In: Forensic Analysis of the Skull. Eds. M. Y. Iscan and R. P. Helmer. Wiley-Liss, New York, 1993, pp. 131159.
Pesce Delfino, V., Lettini, T., and Vacca, E. Heuristic adequacy of Fourier descriptors: Methodologic aspects and applications in morphology. In: Fourier Descriptors and Their Application in Biology. Ed. P. E. Lestrel. Cambridge Press, Cambridge, England, 1997, pp. 250293.
and O. Peschl
J. V. Czarnecki
Wehrwissenschaftiches Institut fuer Materialuntersuchungen
SEM/EDX analysis of the quantitative composition of dental alloys combined with mathematical pattern recognition techniques were used for the identification of dentures in forensic odontology. The statistical interpretation of the data attempts to extract characteristic patterns in the composition of the alloys that can be typical for a certain manufacturer or region.
S. A. Long
d.b.a. I.D. Image Discovery
This poster documents and
pictorially displays the scientific and artistic process of forensic
facial reconstruction (approximation) on a human skull cast with
tissue markers and clay. The skull cast was made from a computer-produced
skull taken from a CT scan of an ancient mummy from Nevada. Final
results as well as an example of variation between a Caucasoid-type
face and a reconstruction on a Native American skull cast are
N. Lynnerup and
University of Copenhagen
In Denmark there has been an increase in armed bank robberies. This has led to an increased demand for facial identification analyses, which usually consist of comparisons between images from surveillance cameras and photographs of alleged perpetrators.
Previously, the image comparisons were performed by police forensic technicians, but recently the forensic odontological and anthropological department began conducting the analyses as a result of the enhanced possibilities for recognizing specific aspects of craniofacial (including mandibular) build, dental occlusion, and general physiognomy of mouth and jaws.
Various raw materials (photographs, videos, and police line-ups) that form the basis of the examinations and methodologies used in these analyses (specific morphological checklists for facial traits, digital imagery, overlays, and on-scene investigations) are discussed. Standardized statements and depositions are also discussed.
M. H. Manhein,
N. E. Barrow, and B. Bassett
Louisiana State University
Baton Rouge, Louisiana
G. A. Listi
New Orleans, Louisiana
At the Forensic Anthropology and Computer Enhancement Services Laboratory (FACES) on the Louisiana State University campus in Baton Rouge, the research-oriented program includes a variety of applications for facial imaging. Not only are the traditional forensic anthropology analyses for local, regional, and national law enforcement agencies performed, but a variety of facial-imaging services are provided: 3D clay facial reconstructions, computer-generated age progressions on missing children and adults, and video and photographic enhancements.
This paper presents selected results of a recent in vivo tissue-depth research project that was conducted over a three-year period (funded in part by a grant from the Louisiana Educational Quality Support Fund). Included are new age-category standards for children and adults. Inter- and intragenerational analyses are also presented.
One aspect of facial imaging includes computer-generated age progressions, which are used in the recovery of missing children and adults. These age progressions are also used by law enforcement to assist in the apprehension of missing fugitives. Examples of successful cases are provided.
Finally, this presentation demonstrates the ability of existing computer software to enhance blurred videos for agencies such as police departments, sheriff's offices, the Federal Bureau of Investigation, and the Bureau of Alcohol, Tobacco and Firearms.
M. K. Marks
Department of Anthropology
University of Tennessee
D. R. Tufano,
E. C. Uberbacher,
R. E. Flanery, V. N. Olman, and Y. Xu
Cognitive and Information Sciences
Computer Science and Mathematics Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee
The objective, wholesale means to expeditiously and economically approximate a believable 3D facial likeness for identification purposes from skeletal, decomposed, fragmented, or mutilated faces is rare. Likewise, the subjective artistic/sculpting skills necessary to create busts from skulls are rare and cost-prohibitive to many law enforcement agencies. Unfortunately, even the best 2D artistic rendition of reality is many times obscure. This research approach in artificial intelligence initiates a quantitative, Internet-based tool kit that can be used to computationally and, therefore, objectively, derive a face from a given skull surface. The collaboration of forensic anthropology, computer science, radiology, and perceptual psychology allows computer-graphic modeling to develop rapid (same day) 3D reconstructions for viewing by family, witnesses, craniofacial experts, or the public.
MRI data were collected from European-American males between 30 and 50 years of age. Statistical methods were derived from two artificial intelligence approaches, case-based reasoning and a neural network, to predict facial surfaces from a skeletal surface. Image segmentation of bone and soft tissue volumes, structural analysis, 3D visualization techniques, and perceptual aspects including use of skin-smoothing algorithms for texture mapping and rotation (e.g., JPEG) to enhance recognition, were attempted. The segmentation algorithms are especially valuable in the reconstruction of fragmented faces.
A face can be produced/predicted from a given skull surface after an electronic file is created using video image capture. The prediction methods using the artificial intelligence approaches decrease the error rate in facial profiles by one half when compared to the traditional clay bust creations. Software and programming development are still in the manufacturing stage for this target (pilot) sample. Future methods will develop and refine facial animation morphing, texture/surface mapping and segmentation, and placement of 3D eyes at various depths in the orbits. This technology will provide any law enforcement agency or medical examiner facility armed with a 2K computer and Netscape a quick, cost-efficient, and reliable means of identifying victims.
This research was funded by a Director's Research and Development Grant at Oak Ridge National Laboratory.
B. A. Martin
Oakland County Sheriff's Department
If a forensic artist has a full-front X-ray, a profile X-ray, and a set of physical criteria for an unknown subject's skull, how would the artist render a subject's image? The following two methods will be evaluated for effectiveness:
- Charting and mapping tissue
depths from the X-rays by overlays or
- Charting and then interacting with team members to enhance overall reproduction and fine-tune important details.
S. Meksi and F.
University of Tirana
The importance of bite mark evidence in adjudicated felony trials such as homicide, sexual crimes, and child abuse in Albania between 1990 and 1999 is demonstrated.
All aspects of bite mark investigation, including the following, are covered:
- Examination of the victim;
- Location and description
of the wound including anatomical area, surface contour, tissue
characteristics, shape, color, type of injury, and size;
- Tissue sample examination
- Photographic documentation;
- Examination of the suspect.
If properly preserved and protected, bite marks can provide an important link between a victim and an assailant.
C. Milner, C.
Wilkinson, and R. Neave
University of Manchester
Manchester, United Kingdom
The adult craniofacial complex undergoes extensive morphological alteration during the adult years of life. Through the postadolescent years, the craniofacial skeleton continues to increase the outward dimensions of the skull within the saggital plane (Behrents 1984). Conventional techniques of facial reconstruction do not account for this process and thus are limited to the prediction of facial characteristics based on skull structure at time of death. With an understanding of the pattern of adult skull growth, it may be possible to predict the morphology of a given skull at other ages during adulthood once the actual age of the skull at time of death is known. Such a technique may be of forensic value when, for example, a skull from a 70-year-old individual could be altered to allow reconstruction at a much younger physical age for identification purposes.
In this work, data was employed from Dr. Rolf Behrents' work of 1984 in order to predict the profile alteration of a 20-year-old female skull over the 45 years subsequent to her death. Behrents' data represent one of the largest and most comprehensive longitudinal studies describing postadolescent craniofacial growth on the basis of serial subject cephalographs (Behrents 1985). These calculations formed the basis for comparative facial reconstruction, which aims to show the possible appearance of the same individual at two time points within his or her lifetime.
The results represent the first attempt to predict the course of craniofacial skeletal growth for an individual through adulthood. Furthermore, the results of this work were developed beyond the cephalometric tracings universally used to represent craniofacial complex growth into a 3D format through facial reconstruction.
Behrents, R. A Treatise on the Continuum of Growth in the Ageing Craniofacial Skeleton. Doctoral Dissertation, University of Michigan, Ann Arbor, Michigan,1984.
Behrents, R. An Atlas of Growth in the Ageing Craniofacial Skeleton. Monograph 18 Craniofacial Growth Series, Center for Human Growth, Ann Arbor, Michigan, 1985.
Prag, J. and Neave, R. Making Faces. Manchester University Press, York, United Kingdom, 1997.
R. A. H. Neave
University of Manchester
York, United Kingdom
The massive proliferation of image surveillance techniques provides a form of evidence that may appear very effective. In many incidences such evidence forms a substantial part of the evidence presented in court. There are, however, few occasions in which the images are of sufficient quality for an unequivocal identification to be made. Security images, whether on conventional film or closed-circuit television, often lack the clarity of definition for fine detail to be seen. The problems are further compounded by the above-eye-level position of the cameras and the tendency of offenders to disguise their appearance.
On many occasions, people see what they want to see rather than what is actually there. It is therefore essential that such evidence is examined with the same rigor as any other forensic evidence. It is also essential that the results be presented to a jury in a form that they can readily comprehend.
This paper explores some of the problems associated with the examination of this type of evidence and the advantages and disadvantages of different techniques. It also will consider the expectations of the investigating authorities and the courts in relation to the degree of reliability of such evidence.
Iscan, M.Y. Introduction of techniques for photographic comparison: Potential and problems. In: Forensic Analysis of the Skull. Wiley, New York, 1993, pp. 5770.
Linney, A. and Coombes, A. M. Computer Modelling of Facial Form: Craniofacial Identification in Forensic Medicine. Arnold, London, 1998, pp.189199.
G. L. Nusse
Guild of Natural Science Illustration
Mill Valley, California
The skull of the Taung Child, Australopithecus africanus, was facially reconstructed. Contrasting reconstructive techniques will be displayed: one using muscle suture lines of the skull and the other using standard tissue-depth markers. Both reconstructions are based on a casting of the original skull found by Dr. Raymond Dart in 1924. This cast was obtained courtesy of the Institute of Human Origins in Tempe, Arizona.
V. M. Phillips
Oral and Dental Teaching Hospital of the University of Stellenbosch
Cape Town, South Africa
The identification of human remains is of paramount importance for legal and humane reasons. The reconstruction of the facial features of an individual onto the skull is a blending of the scientific and the artistic skills of the sculptor. This method is usually a last resort to identify the skeletal remains of an unidentified person, and it suffers from an ongoing skepticism caused by the advent of the personal computer and modern software technology. There are numerous techniques to sculpture a face onto the skull, all of which rely on the reproduction of a potentially recognizable face using the published soft tissue thicknesses for different racial groups.
Three incidents in which facial sculpturing was used to identify victims of unnatural deaths are reported. The first was the identification of the remains of a suicide victim found on the summit of Table Mountain in Cape Town, South Africa. The second was the identification of victims of a political assassination in which four young men were shot in the back of the head. The final identification was that of a young girl whose body was found in a shallow grave in Cape Town.
The sculpturing method of facial reconstruction has merit and yields remarkable results, including the gratitude of the relatives of the identified victim and the satisfaction of the forensic anthropologist.
National Institute of Public Health
Prague, Czech Republic
D. H. Ubelaker
National Museum of Natural History, Smithsonian Institution
Reconstructing the nose shape from remains of the skull is problematic. Some authors considered it impossible. The famous laboratory of M. M. Gerasimov in Moscow devoted much experimental effort to solving this problem. One of the authors (M. Prokopec) was trained by Gerasimov's successor, Galina Lebedinskaya, in the method developed and used in the laboratory. This method reconstructs the nose shape on the basis of a skull, provided the nasal bones and the midface of the cranium remained intact.
For this study, four well-preserved skulls (two male and two female) from a Slavonic cemetery at Rajhrad, Czech Republic, dating from the eleventh century were found. The skulls were used to perform 2D facial reconstructions and to demonstrate the method of nose shape reconstruction.
A perfect drawing of each skull was made from its profile with a dioptrograph. These drawings displayed the following details: maxilla, os nasale, contours of eye sockets, os zygomaticum, processus zygomaticus, and all sutures.
- A line (A) is drawn through
the points nasion and prosthion.
- A parallel line (B) is then
drawn, intersecting the foremost point on the nasal bone.
- Four or five lines (C, D,
E, F, G) in equal distances between each other are drawn perpendicular
to line B on its section from the very tip of the nasal bone
to the base of the apertura piriformis. Each of these lines crosses
line B and has an inner and outer section. The outer section
starts at line B.
- The distance from the rim
of the apertura piriformis to line B on line C (inner section)
is marked on its outer part by a dot. The same is done on each
of the lines D, E, F, and G.
- The dots on the outer sections of lines C, D, E, F, and G were connected with a curve, and the mean thickness of the skin and fat layers at this area were added.
This procedure provides the most probable contour of the nose of a person's skull.
Dr. Lebedinskaya also described
the most probable position of the eyeball and eyelid, the size
and angle of the outer ear, the midline between the lips, the
position of the mouth corner, and the form of the chin. These
descriptions were considered in the four examples of 2D reconstructions.
Hair structures in both sexes and beards in the men were adjusted
freely according to the authors' interpretations, presuming that
such styles were worn by the Old Slavs in the eleventh century.
FORENSIC SCIENCE COMMUNICATIONS OCTOBER 2000 VOLUME 2 NUMBER 4