Petra Urbanová
New Dimensions of
Forensic Anthropology
NewDimensionsofForensicAnthropology
Petra Urbanová
Masaryk University
Faculty of Science
New Dimensions of
Forensic Anthropology
Brno 2016
Cover design by Mikoláš Jurda
Laboratory of Morphology and Forensic Anthropology
Masaryk University, Faculty of Science, Brno, 2016
Preface 									
Acknowledgements 								
List of co-authors 								
1. Forensic Osteology					
1.1.	Advanced Methods in 3-D Craniofacial Morphological Analysis		
1.2.	Complexity of Assessing Migrant Death Place of Origin	
1.3.	Testing the Reliability of Software Tools in Sex and Ancestry Estimation in
	 a Multi-Ancestral Brazilian Sample
1.4.	The Morphology of Human Hyoid Bone in Relation to Sex, Age and Body 		
	 Proportions
1.5.	The Asymmetry and Modularity of the Hyoid Bone
1.6.	What Is the Appropriate Approach in Sex Determination of Hyoid Bones?
1.7.	Two-Dimensional Wavelet Analysis of Supraorbital Margins of the
	 Human Skull for Characterizing Sexual Dimorphism
2. Image-based Identification of the Living
2.1.	3D Virtual Model Database of Human Faces: Applications in
	 Anthropology and Forensic Sciences
2.2.	Performance of Distance-based Matching Algorithms in 3D Facial
	 Identification
2.3.	Generating Various Composite Human Faces from Real 3D Facial Images
3. Advanced Digital Techniques in Forensic Postmortem 		
	 Examination
3.1.	Testing Photogrammetry-based Techniques for Three-dimensional Surface
	 Documentation in Forensic Pathology
3.2.	Tethered Digital Photography with Built-in Wi-Fi Memory Cards Brings 		
	 Benefits to the Environment of an Autopsy Room
4. Forensic Skeletal Traumatology
4.1.	Agenesis of the Superior Cornua of the Thyroid Cartilage. A Rare Variant of
	 Medicolegal Importance
4.2.	Virtual Approach to the Assessment of Skeletal Injuries in Human Skeletal
	 Remains of Forensic Importance
CONTENTS
V
VII
IX
X
1
7
39
61
73
91
103
113
123
129
135
155
173
177
189
195
201
207
In the eyes of many, forensic anthropology continues to be taken for an appendix of physical anthropology,
whose focus is on identification of skeletal remains within a context of criminal investigation.
Under this concept a forensic anthropologist is expected to find definite answers to three fundamental
questions, asked upon the discovery of skeletal remains – how long has it been since the person died?
Can the remains be identified? Is there any evidence of skeletal trauma, which can pinpoint the cause
of death? This is, somewhat, skewed perception of forensic anthropology, fueled, among others, by an
inaccurate portrait of anthropologists in media and pop culture, presented shamelessly as the “techsavvy
all-knowing bone collectors”.
There are only a few practicing forensic anthropologists in our country. Having the privilege to be one
of them has given me first-hand experiences to the eclectic nature of forensic casework and has provided
me with an overview to recent shifts occurring in the intellectual foundation of the field.
The focus of this volume is to offer a better understanding of forensic anthropologist’s work and to emphasize
the importance of balancing anthropologist’s skills with those adopted from other disciplines.
Taking up a forensic case an anthropologist must possess the knowledge of biological (and not only
biological) principles that goes far beyond a perfect understanding of human osteology, because today’s
forensic casework no longer demands a conventional methodological basis. It calls for multi-disciplinary
approaches. Diverse types of evidence submitted for forensic examination and the booming
development in imaging and computational technologies have added additional layers of scientific
knowledge that has become indispensable in conducting a forensic examination. Throughout the progress
in imaging technologies, particularly three-dimensional techniques, forensic sciences have become
more dependent on numerical and visual data processing. This has introduced new dimensions,
both real and theoretical, into the framework of forensic anthropology.
The volume, itself, is a collection of peer-reviewed scientific studies published (with one exception of a
submitted manuscript) between years 2013 and 2016. The studies are structured into four main chapters.
Each chapter opens with an introductory section providing an overview to a specific area of forensic
anthropology casework or research and is followed by a short commentary on the published papers
featured inside.
A recurring theme to the papers composing the volume is the employment of 3D technologies, modulated
to be applicable to forensic casework. Over the last decade, the progress in 3D systems has launched
multiple novel applications in forensics. Examination of fragmentary skeletal remains started relying
on advanced geometric variables and virtual approaches (Chapter 1). Identification of living persons
has turned to 3D face and motion recognition (Chapter 2). Traditional invasive “scalpel and scissors”
dead body examination started being accompanied by non-invasive diagnostics grounded on
full-body CT scans and a robotic assistance (Chapter 3 and 4). Computer-aided reconstructions of an
incident featuring digital evidence of skeletal injuries have been created to aid investigations of violent
deaths (Chapter 4). These techniques are rapidly becoming part of experts’ life and are transforming
the manner in which forensic casework is conducted.
The primary purpose for this volume was to straighten the public misconception of forensic anthropologists
and to offer a commentary on the latest development in the field. Hopefully, I have succeeded
at least in one of these goals.
Brno, August 2016
							 Petra Urbanová
PREFACE
VII
The most thankworthy collaborators are acknowledged as co-authors on the next page and again
throughout the volume. Therefore, I wish to dedicate this page specifically to those whom I had not
had the pleasure to collaborate with on any of the featured studies. Yet, their professionalism, skills,
erudition and wisdom have been inspirations, which transpire through the publications.
To dr. Hana Eliášová, Institute of Criminalistics Prague, my ever so patient and understanding mentor
in forensic casework;
To employees of the Departments of Forensic Medicine in Brno (Masaryk University) and Hradec
Králové (Charles University) for being excellent “partners in crime” willing to participate in our experiments
with imaging technologies in forensic medicine and forensic pathology;
To prof. dr. Edna Sadayo Miazato Iwamura, Department of Pathology, Federal University of São Paulo,
and dr. Paulo Eduardo Miamoto Dias for their hospitality, support and assistance in our ongoing
Czech-Brazilian collaboration;
To team members and students participating in the FIDENTIS Project at Masaryk University for their
dedication, insights, effort and for keeping up the good work;
To co-workers and students at the Laboratory of Morphology and Forensic Anthropology for
showing up to work every day and only occasionally giving a headache.
Many thanks to the Fulbright commission for awarding me with the Fulbright Scholar grant. They say
that being a Fulbrighter is for life…and I proudly add that rightly so. The U.S. milieu has made a great
impact on professional decisions and has shaped my interests in the field. Also, special thanks to
the Council for International Exchange of Scholars and the Outreach Lecturing Fund sponsored by
the U.S. Department of State for providing me with the funding to visit forensic laboratories in Oahu,
Hawaii, USA.
The published work presented in this volume has been undertaken with the financial assistance of
Masaryk University. Together with the rest of the Laboratory of Morphology and Forensic Anthropology,
I am grateful that they keep funding our research.
ACKNOWLEDGEMENTS
IX
Roberto M. Cesar, Jr.
Department of Science Computer, Institute of Mathematics and Statistics, University
of São Paulo, São Paulo Brazil
Zuzana Ferková
Human Computer Interaction Laboratory, Faculty of Informatics, Masaryk University,
Brno, Czech Republic
Petr Hejna
Department of Forensic Medicine, Faculty of Medicine, Charles University in Prague,
Prague; University Hospital Hradec Králové, Hradec Králové, Czech Republic
Igor Chalás
Human Computer Interaction Laboratory, Faculty of Informatics, Masaryk University,
Brno, Czech Republic
Marie Jandová
Laboratory of Morphology and Forensic Anthropology, Department of Anthropology,
Faculty of Science, Masaryk University, Brno, Czech Republic
Martin Janík
Institute of Legal Medicine and Medico-legal Expertises, Jessenius Faculty of Medicine,
Comenius University, Martin, Slovakia
Chelsey A. Juarez
North Carolina State University, Raleigh, North Carolina, USA
Mikoláš Jurda
Laboratory of Morphology and Forensic Anthropology, Department of Anthropology,
Faculty of Science, Masaryk University, Brno, Czech Republic
Vojtěch Juřík
Department of Psychology, Masaryk University, Brno, Czech Republic
Zuzana Kotulanová
Laboratory of Morphology and Forensic Anthropology, Department of Anthropology,
Faculty of Science, Masaryk University, Brno, Czech Republic
Barbora Kozlíková
Human Computer Interaction Laboratory, Faculty of Informatics, Masaryk University,
Brno, Czech Republic
Maria-Ines Nogueira
Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil
List of co-authors (alphabetically)
X
Sílvia C. D. Pinto
Department of Science Computer, Institute of Mathematics and Statistics, University of
São Paulo, São Paulo Brazil
Ann H. Ross
North Carolina State University, Raleigh, North Carolina, USA
Jiří Sochor
Human Computer Interaction Laboratory, Faculty of Informatics, Masaryk University,
Brno, Czech Republic
Miroslav Šafr
Department of Forensic Medicine, Faculty of Medicine, Charles University in Prague,
Prague; University Hospital Hradec Králové, Hradec Králové, Czech Republic
Lenka Zátopková
Department of Forensic Medicine, Faculty of Medicine, Charles University in Prague,
Prague; University Hospital Hradec Králové, Hradec Králové, Czech Republic
XI
Forensic Osteology
Chapter 1
3
Chapter 1: Forensic Osteology
Examination of skeletal remains of unknown origin continues to prevail as the primary task for forensic
anthropologists despite the regional diversity in formal training and personal preferences for daily
casework. As a result, the main effort is given to tasks involving identification of fully or partially skeletonized
human remains. The routine osteological procedure starts with assessment of ancestry, biological
sex, age-at-death and stature and carries on to examination of unique skeletal or dental traits to
match or eliminate the person’s identity. Basic, yet pertinent methods to carry out these tasks include
visual assessment (e.g., morphoscopic or qualitative traits) and osteometrics (e.g., anthropometrics,
craniometrics, quantitative traits).
In the last 30 years, a substantial body of work has been achieved in order to evolve analytical and diagnostic
tools used in forensic osteology with a primary goal to move away from the subjectivity of visual
comparison and the triviality of caliper-based measurements. Additionally, another concern has
shaped the manners in which a forensic osteological expertise is nowadays executed. In mid-1990’s the
admissibility of many traditional osteological techniques in courts started being questioned and the
criticism particularly intensified upon passing of the Daubert standards. Named after the ruling in the
US case Daubert vs. Merrell Dow Pharmaceuticals, the standards state that scientific evidence admitted
in court must be collected with known error rates. It also must reflect the scientific knowledge accepted
among peers following dissemination in peer-reviewed scientific publications. Therefore, starting
with the new millennium most of the research in forensic osteology has aimed at quantifying probability
rates upon which the parameters retrieved from skeletal remains are estimated. The major
goal has been to provide forensic experts with a more complex methodological basis that would fulfill
the legal criteria and simultaneously would enhance the precision and replicability of used standards.
This has logically placed higher demands on anthropologist’s operating and analytical skills as well as
technical lab equipment.
The state-of-the-art of forensic osteology is linked to two principal trends of systematic development.
The first of them has been tied to the gradual employment of multivariate statistical algorithms combined
with large sample sizes. This has allowed making use of the exploitative role of statistics as opposed
to the previous descriptive and summarizing function. It also allowed interlacing multiple predictive
skeletal features into a single complex descriptor. Applicable to both quantitative and qualitative
variables, multivariate classification and predictive algorithms have become integral and irreplaceable
parts of daily anthropological practice enabling to allocate an unknown individual into a demographic
category (e.g., ancestral groups, biological sex etc.) with a given probability score or to estimate
an individualizing criterion with a predicted error (e.g., stature). The field has also witnessed a tendency
towards more progressive employment of highly sophisticated approaches based on Bayesian
statistics or machine learning algorithms in form of classification trees or neural networks.
As the second research avenue of forensic osteology became focused on maximizing the amount and
complexity of processed skeletal data, it became evident that the traditional approach utilizing trivial
metric variables taken with measuring instruments cannot keep up with such demands. By contrast,
the substantial progress accomplished in digital technologies permitted the improvement when linear
or angular measurements started being replaced with geometry-derived variables (e.g., diameters,
three-dimensional landmarks, radial coordinates, non-linear distances, outlines or curves). These variables
processed by methods of geometric morphometrics or statistical shape analysis have been
shown to enhance significantly assessments carried out in the course of examination of skeletal rema-
ins.
The most recent inspiration for the forensic practice has derived from parallel development made in
three-dimensional acquisition technologies and computation performances. This increased availability
and affordability of a wide range of recording hardware featuring contact and contactless digitizers,
optical and laser scanners or medical imaging units. It also allowed collecting high-resolution 3D
digital models of skeletal remains. Nowadays, 3D digital models of skeletal remains form sizable datasets
employable as references on intra-population and inter-population skeletal variation. They are
4
Chapter 1: Forensic Osteology
taken for truthful replicas protecting the integrity of perishable physical evidence. They are even preferred
as easily displayable alternatives to the real evidence when presented to investigators or jury
members. And more importantly, they are used as the aforementioned sources of abundant quantitative
and qualitative skeletal data.
Despite the outlined progress, integrating advanced methods into an everyday routine can be challenging.
The novel approaches ask for larger learning curves and although a colorful spectrum of advanced
computer-aided techniques is available, only a few of them are tailor-made to accommodate the
level of computer literacy commonly encountered in forensic anthropologists. For experts who have
difficulty keeping up with technological novelties due to caseloads it means that mastering additional
necessary skills may be virtually beyond reach. In order to overcome these challenges team efforts
are being extended to simplify scientific knowledge, to ease computational requirements and to offer
experts user-friendly computer environments to treat morphological data retrieved from skeletal re-
mains.
This chapter contains seven published studies falling under the rubric of forensic osteology. They aim
to illustrate author’s principal contributions to the development in the field. The first two titles “Advanced
Methods in 3-D Craniofacial Morphological Analysis” and “Complexity of Assessing Migrant Death
Place of Origin” are book chapters published in “Biological Distance Analysis. Forensic and Bioarchaeological
Perspectives” edited by Marin A. Pilloud and Joseph T. Hefner. The chapters were co-authored
by dr. Ann H. Ross and dr. Chelsey Juarez, both currently affiliated with North Carolina State University,
Raleigh, USA and originated in current needs to re-valuate and update methods used to estimate
ancestral origin from skeletal remains, especially if used to allocate an individual within a population
of a complex and admixed background (e.g., illegal Hispanic/Latin immigrants in the United States or
European and Admixed Brazilians in Brazil).
Together with the next paper “Testing the Reliability of Software Tools in Sex and Ancestry Estimation
in a Multi-Ancestral Brazilian Sample”, the book chapters form a trilogy which provides insight into
advanced 3D craniofacial analysis as performed in the course of forensic casework. The papers emphasize
on the usage of computer-aided diagnostic tools within various world populations. They illustrate
that three-dimensional data acquisition hardware combined with methods of advanced data processing
is capable of boosting diagnostic powers and subsequently increases levels of accuracy. The results
also send a warning to the forensic community that the advanced techniques should be applied
with extreme caution because a number of missteps in form of an inadequate reference sample or an
improper handling of a technique can seriously skew acquired results.
The trio on craniofacial variation is followed by another set of studies tied by a central theme – morphological
variations of human hyoid bones. As part of skull, hyoid bone is an important, yet sometimes
overlooked, element of human skeleton. The three studies demonstrate that hyoid bone represents a
valuable source of information on individual’s sex (“The morphology of human hyoid bone in relation to
sex, age and body proportions”) as well as developmental issues (“The Asymmetry and Modularity of the
Hyoid Bone”). More importantly, it plays a key role in forensic cases where foul play is suspected (see
Chapter 4). Making hyoid bone a model skeletal element, the third of the hyoid-themed papers (“What
is the appropriate approach in sex determination of hyoid bones?”) illustrates the range of multivariate
statistical methods on which a predictive model to estimate identity-related skeletal indicators can be
established.
The chapter concludes with a paper entitled “Two-Dimensional Wavelet Analysis of Supraorbital Margins
of the Human Skull for Characterizing Sexual Dimorphism”, published with Sílvia Pinto and Roberto
Cesar, Jr., author’s long-time Brazilian collaborators affiliated with University of São Paulo. The paper
follows biologists’ (or author’s for that matter) long-standing interests in quantifying biological
shape variations (cross-sections to the supraorbital margin in this case) using progressive mathematical
principles. In summary, the paper was written with objectives to overcome inconveniences of the
5
Chapter 1: Forensic Osteology
Fourier analysis, a relatively widespread shape quantifying technique whose core is the Fourier transform
or harmonic decomposition. While the method excels in quantifying global shapes, it is less successful
in specifying localized morphological abnormalities. By employing a decomposition technique
grounded on a series of multiresolution wavelet functions (i.e., wavelet analysis), capable of specifying
both global as well as local shape variations, the results show that the newly proposed approach, albeit
elaborate, may represent a way to improve performances of sex estimation methods if incorporated
as an integral part of anthropological methodology.
7
Chapter 1: Forensic Osteology
1.1. Advanced Methods in
3-D Craniofacial Morphological
Analysis
by Petra Urbanová, Ann. H. Ross
published in “Biological Distance Analysis. Forensic and
Bioarchaeological Perspectives” edited by
Marin A. Pilloud and Joseph T. Hefner
39
Chapter 1: Forensic Osteology
1.2. Complexity of Assessing
Migrant Death Place of Origin
by Ann. H. Ross, Chelsey A. Juarez,
Petra Urbanová
published in “Biological Distance Analysis. Forensic and
Bioarchaeological Perspectives” edited by
Marin A. Pilloud and Joseph T. Hefner
61
1.3. Testing the Reliability
of Software Tools in Sex and
Ancestry Estimation in a MultiAncestral
Brazilian Sample
by Petra Urbanová, Ann H. Ross, Mikoláš Jurda,
Maria-Ines Nogueira
published in Legal Medicine 16 (2014)
Chapter 1: Forensic Osteology
73
1.4. The Morphology of Human
Hyoid Bone in Relation to Sex, Age
and Body Proportions
by Petra Urbanová, Petr Hejna, Lenka Zátopková,
Miroslav Šafr
published in HOMO - Journal of Comparative Human
Biology 64 (2013)
Chapter 1: Forensic Osteology
91
1.5. The Asymmetry and
Modularity of the Hyoid Bone
by Petra Urbanová, Petr Hejna, Lenka Zátopková,
Miroslav Šafr
published in International Journal of Morphology 32
(2014)
Chapter 1: Forensic Osteology
103
1.6. What Is the Appropriate
Approach in Sex Determination of
Hyoid Bones?
by Petra Urbanová, Petr Hejna, Lenka Zátopková,
Miroslav Šafr
published in Journal of Forensic and Legal Medicine 20
(2013)
Chapter 1: Forensic Osteology
113
1.7. Two-Dimensional Wavelet
Analysis of Supraorbital Margins
of the Human Skull for
Characterizing Sexual
Dimorphism
by Sílvia C. D. Pinto, Petra Urbanová,
Roberto M. Cesar, Jr.
published in IEEE Transactions on Information
Forensics and Security 11/7 (2016)
Chapter 1: Forensic Osteology
Image-based Identification
of the Living
Chapter 2
125
Chapter 2: Image-based Identification of the Living
Identification of living persons represents a specific, sometimes marginalized, area of forensic anthropology,
nowadays linked almost exclusively to digital or digitized image evidence (photographs,
X-rays, web-cam records etc.). Yet, images are valuable and nowadays omnipresent conveyors of individual
and group specific features that point out the person’s identity. In general, identification of
living persons is conducted by matching DNA samples, imprints of various body parts (fingerprints,
handprints, footprints, lip imprints, ear imprints), dental or medical records or physical appearance.
Unfortunately, only some of these individualizing characteristics are observable or retrievable from
image evidence secured at crime scenes. Of the few applicable, facial features, body proportions, gait
dynamics and occasionally morphology of other body parts have been shown as the most appropriate
for image-based identification.
The crucial element in image-based identification tasks is the quality of provided evidence. Standardized
portrait photographs taken under controlled conditions are being submitted as evidence in cases
where person’s real identity is being questioned, such as false passports, frauds, money laundering,
drug smuggling. In addition, three-part (or two-part) police photographs, otherwise known as mugshots
or booking photos, have been taken as photographic evidence while in police custody since the
Alphonse Bertillon’s era. In most cases, however, images processed for identification purposes originate
in uncontrolled, real-life situations, e.g., family snapshots, or in recordings of a crime, e.g., CCTV
surveillance videos capturing a bank robbery.
Ideally, in order to extract and classify identity-specific body traits presented on images semi-automatic
or automatic algorithms adopted from the fields of biometrics, signal and image processing or
computer vision are employed. These algorithms, used mostly in commercial security systems and
grouped under verification/authentication procedures and facial recognition, perform either one-to-one
or one-to-many matching tasks. Yet, a universal use of these algorithms in forensics is problematic.
Although various law-enforcement agencies have been successfully using automatic face recognition
systems for screening and large-scale database inquiries, other tasks, such as negative identifications
(i.e., ability to eliminate a person among many) or positive identification (i.e., ability to match a person
among many) are proved more difficult to achieve.
It is a well-recognized issue that the tasks of personal identification (same-person comparison, inter-person
comparison) are obstructed by the fact that image-based forensic evidence is presented
for examination in very poor quality, with non-standard framing, viewpoints or optical distortion. In
addition, perpetrator’s behavior, occlusion or face expressions can pose additional inconveniences.
In result, these images are rarely suitable for automatic or semi-automatic face recognition systems
because discriminatory powers of individualizing features are diminished by technical noise and intra-individual
variations due to an unstable relationship between a capturing system and a captured
subject.
One of the few manners how to approach this kind of evidence is by means of anthropological examination.
Generally, forensic anthropologists are accustomed to address tasks of similarity of identity-coding
features depicted in images by one of three universal strategies: visual (morphological)
assessment, somatometrics (anthropometrics, photo-anthropometry) or superimposition. By visual
assessment, a specific body feature (e.g., facial components) is assigned into categories according to a
standardized classification system. Various systems for classifying facial features (e.g., the facial form,
nose tip shape, septum tilt etc.), body characteristics (e.g., somatotypes, posture) or descriptive parameters
of a gait cycle (e.g., out-toed gait, drop foot, waddling gait etc.) have been developed. The metric
approach, in contrast, relies on standardized measurements taken for a given body part (e.g., head
length/width, arm length, step length) or a combination of them (e.g., height, cadence of gait). Alternatively,
advanced quantitative techniques based on 2D or 3D spatial landmarks or outlines (silhouettes)
are employed in casework of a more progressive nature. Ultimately, superimposition techniques place
one set of probe images on top of target images and adjust them to make a visual comparison.
Unfortunately for forensic anthropologists, none of the universal strategies is universally applicable.
126
Chapter 2: Image-based Identification of the Living
The visual approach has been scrutinized as largely subjective and strained by a relatively high intraobserver
as well as inter-observer error rate. The metrics are known to be limited among others by
image quality, camera angles and potential inconsistencies throughout person’s lifetime (e.g., height,
gait pattern, facial proportions) and the superimposition does not provide any measurable declaration
of concordance between two superimposed images, nor does it allow assigning a probability score to
a positive or negative outcome.
The most noticeable development in the area of forensic image-based identification in the last ten years
has been linked to the development of 3D sensors. Nowadays, 3D images of living persons can be
captured by a variety of in-lab devices. Dynamic body features – gait or hand gestures – can be acquired
by precise motion tracking systems or even by systems designed to accessorize personal electronic
devices such as tablets, cell phones or game consoles. This has allowed recording realistic images
featuring depth information otherwise lost in the traditional images. The depth information has been
largely advertised as capable of compensating for at least some of the key shortcomings associated
with the traditional image-based methodology, primarily by offering 2D to 3D and 3D to 3D digital record
matching of human body. Although this claim is not without its merits, 3D capturing devices are
strained by their own technical limits (e.g., inability to record obscured regions and dynamic or reflective
body features as such as hair, facial expression etc.), which can interfere with identification tasks
and which are yet to be resolved.
The chapter includes three published papers that demonstrate author’s contributions to the subject
of forensic facial identification. All of them emerged from the FIDENTIS Project at Masaryk University.
The FIDENTIS Project (the acronym stands for Forensic Identification System) is a multi-disciplinary
research group, which teamed up anthropologists with computer scientists in order to join efforts
in researching issues of 3D facial identification. The project benefits from top-tier 3D recording equipment
available at collaborators’ workplaces and capable of collecting various types of 3D datasets
(e.g., Vectra M1 optical scanner, Vectra XT half-body system, Vectra H1 scanner, Facial Motion Capture
System, Microsoft Kinect). To date, the project has given origin to two main achievements: FIDENTIS
3D Face Database (F3D-FD) and FIDENTIS Analyst software (both available via project web pages
at www.fidentis.cz).
The first paper titled “3D Virtual Model Database of Human Faces: Applications in Anthropology and
Forensic Sciences” provides an overview of the FIDENTIS 3D Face Database. The database represents
one of the largest datasets of 3D facial scans currently available worldwide. It composes of single-scan
entries and a smaller number of multi-scan entries varying in acquisition conditions. In 2014, which
coincides with the date of publication, the database contained approximately 1,700 3D face scans. To
date, 3D face scans and additional demographic data from over 2,300 participants (both adult and
sub-adults) have been collected and listed in the database. The primary purpose of the database is to
serve as a stand-alone reference sample encompassing the maximum possible range of inter- or intraindividual
variations of facial features.
The second paper “Performance of distance-based matching algorithms in 3D facial identification” provides
an introduction to functionalities incorporated into FIDENTIS Analyst program. FIDENTIS Analyst
is a target-orientated computer interface designed for processing 3D images of human faces. Processing
two or multiple 3D images is a challenging and computationally expensive task. In order to compare
3D images a forensic expert must possess sufficient familiarity with a number of 3D data processing
techniques, such as 3D alignment, mesh undersampling, color-mapping etc. The software is being
developed to facilitate many of these necessary tasks. For 3D facial images it enables a variety of
morphological analyses – a 3D facial composite construction, automated landmark localization and
face-to-face comparison in one-to-one, one-to-many and batch processing modes. The featured study
tests performances of computation algorithms comprising 3D-to-3D image matching procedures. Spe-
127
Chapter 2: Image-based Identification of the Living
cifically, it confronts identification rates based on manually collected 3D facial landmarks and automatic
processing of 3D meshes. It is noteworthy that the testing was conducted on a dataset of 3D faces
sampled from the FIDENTIS 3D Face Database.
The last included paper represents a large-scale collaborative study entitled “Generating various composite
human faces from real 3D facial images”, which hints on another aspects of image-based forensic
expertise – image composites. The study extends the traditional approach to the facial composites by
introducing algorithms for creating 3D facial images based on real life data (3D facial components)
extracted from the FIDENTIS Database. Although the paper was targeted at the non-forensic scientific
community (it was meant for the gaming industry), the goal of presenting computation principles for
generating composite human faces coincides with that of the forensics-oriented research. A two-fold
user study was added to the descriptive sections of the paper to test a hypothesis whether the perception
of human faces had shifted once one or more facial components had been modified. The outcome
was confronted against those of a morphological analysis quantifying the extent of such facial alterations.
The results suggest that to an extent the proposed automated algorithms are capable of producing
realistic 3D face composites that are far from being caricatures and morphologically fall within the
range of observed facial variation.
129
Chapter 2: Image-based Identification of the Living
2.1. 3D Virtual Model Database of
Human Faces: Applications in
Anthropology and Forensic
Sciences
by Zuzana Kotulanová, Igor Chalás,
Petra Urbanová
published in The Dolní Věstonice Studies, vol. 20 Mikulov
Anthropology Meeting (2014)
135
Chapter 2: Image-based Identification of the Living
2.2. Performance of Distancebased
Matching Algorithms in 3D
Facial Identification
by Petra Urbanová
published in Egyptian Journal of Forensic Sciences 6
(2016)
155
Chapter 2: Image-based Identification of the Living
2.3. Generating Various
Composite Human Faces from
Real 3D Facial Images
by Igor Chalás, Petra Urbanová, Vojtěch Juřík,
Zuzana Ferková, Marie Jandová, Jiří Sochor,
Barbora Kozlíková
published in The Visual Computer 32
(2016)
Advanced Digital Techniques
in Forensic Postmortem
Examination
Chapter 3
175
Chapter 3: Advanced Digital Techniques in Forensic Postmortem Examination
The gold standard of forensic documentation is conventional two-dimensional photography. Photography
is the very first technique employed for documenting forensic scenes, dead bodies or skeletal
remains on sites as well as throughout the entire process of in-lab postmortem examination (external
assessment and internal 3-cavity autopsy). The core disadvantage of traditional photographs is the
lack of depth information, which inherently reduces spatial relationships among the recorded objects
to a two-dimensional level. This limits experts’ abilities to interlink contextual information drawn from
on-site photography with autopsy findings. It also diminishes chances to revisit original findings made
during postmortem examinations.
In light of the recent achievements made in digital imaging technologies, several projects have invested
into establishing three-dimensional postmortem documentation protocols by utilizing a variety of
3D acquisition hardware ranging from haptic devices, to close-range photogrammetry, to robotic surface
scanning. It has been well demonstrated that these techniques bring numerous benefits in decision-making
conducted inside as well as outside autopsy rooms or forensic laboratories. They manage
to produce high-quality digital copies of real-life objects, which can be examined from infinite viewpoints,
edited, revised, measured, shared by an unlimited number of practitioners and brought back to
the physical world by the technology of 3D printing. If combined with high-resolution surface color information
(i.e., texture) they also provide information about time-sensitive coloring, spots, or subtle
external interferences important in diagnostics.
Three-dimensional postmortem documentation taken for an alternative to traditional photography is,
however, part of a larger initiative oriented at establishing imaging technologies for the purpose of internal
postmortem examination – forensic autopsy. Optical surface models are mere digital shells lacking,
by definition, any information going underneath the surface. By contrast, medical or technical
imaging modalities such as computed tomography, cone-beam CT, microCT or MRI generate cross-sectional
(volume) data, which have the capacity to produce 3D model of any internal structure. Altogether
the approach composed of non-invasive or minimally invasive techniques in postmortem examination
is referred to as the virtual, also touch-free or scalpel-free autopsy.
For forensic anthropologists requested to assist a forensic pathologist in cases of assessment of skeletal
injuries or identity confirmation, the image-based approach represents a welcome alternative to
the time-consuming maceration and soft-tissue processing. 3D digital models have been proven valuable
as virtual substitutes for physical bones in sex and ancestry assessment, age estimation or facial
reconstructions. CT scans of arms and legs can uncover individual, sex or age-related markers – degree
of ossification (epiphyseal fusion, hyoid bone, and thyroid cartilage), ossified or calcified elements
(microtrauma, healed injuries, atherosclerotic lesions). In traditional autopsies, where only a
portion of the body is examined, mostly for time-saving, economic or financial reasons or reasons of
piety, these findings can easily go unnoticed. Furthermore, CT images of gross bone morphology, skeletal
non-metric traits, surgical evidence, implants, endoprostheses, otherwise known as “morphological
fingerprints” are helpful in identification efforts. Freely movable and rotatable CT-generated 3D
models are particularly beneficial if a comparison with antemortem medical evidence is carried out.
As postmortem examination may evoke unwanted connotations image-based visual data also offer a
psychological release when graphic forensic evidence is presented and the cause of death is explained
to the police, jury members or next of kin.
Conducting image-based postmortem examination comprises a series of non-trivial consecutive tasks
and by nature has to be collaborative and multidisciplinary. Nowadays, a strong voice can be heard
calling for large-scale multi-centers that would be able to take on such responsibility. Still, the number
of interdisciplinary centers worldwide is scarce and it is not expected to grow drastically. Moreover,
due to their heavy caseload, forensic practitioners are rarely inclined towards experimenting with novel
technologies and incorporating experimental approaches into everyday workflow. Yet, it has been
shown that a symbiosis of sorts can be achieved on a smaller scale, particularly through collaboration
176
Chapter 3: Advanced Digital Techniques in Forensic Postmortem Examination
with other experts, such as forensic anthropologists who are generally lighter on daily casework, more
research-oriented and yet competent to assist at postmortem examination.
This chapter contains two studies, which touch the subject of implementing technical and technological
novelties into the routine workflow of forensic postmortem examination. Both of them also highlight
the fruitfulness of complementing pathologist’s skillsets with those of a trained forensic anthropologist.
The first paper entitled “Testing photogrammetry-based techniques for three-dimensional surface
documentation in forensic pathology” marks the first attempt to employ techniques of optical surface
scanning and single camera close-range photogrammetry as stand-alone alternatives to conventional
forensic photography conducted inside an autopsy room. The acquired results particularly underline
the fact that three-dimensional documentation at autopsies does not have to be costly luxury
requiring top-notch devices, but it can be carried out with standard lab equipment.
Thematically, the second paper “Tethered digital photography with built in Wi-Fi memory cards brings
benefits to the environment of an autopsy room” represents a step back from the 3D imaging techniques
presented in the previous study by focusing on traditional photographic techniques. However, alike its
predecessor, the paper introduces a sophisticated strategy that facilitates expert’s routine workflow
and increases the efficiency of teamwork inside an autopsy room or forensic laboratory. Specifically,
the paper points out the benefits of using built-in wireless fidelity memory cards allowing cable-free
transmission of images, their real-time visualization and instant quality control.
177
Chapter 3: Advanced Digital Techniques in Forensic Postmortem Examination
3.1. Testing Photogrammetrybased
Techniques for
Three-dimensional Surface
Documentation in
Forensic Pathology
by Petra Urbanová, Petr Hejna,
Mikoláš Jurda
published in Forensic Science International 250
(2015)
189
Chapter 3: Advanced Digital Techniques in Forensic Postmortem Examination
3.2. Tethered Digital Photography
with Built-in Wi-Fi Memory Cards
Brings Benefits to the
Environment of an Autopsy Room
by Petr Hejna, Martin Janík,
Petra Urbanová
published in Romanian Journal of Legal Medicine 23
(2015)
Forensic Skeletal
Traumatology
Chapter 4
197
Chapter 4: Forensic Skeletal Traumatology
The main purpose of forensic (medicolegal) autopsy is to determine whether death occurred as a result
of natural causes, suicide, or to confirm or reject foul play when suspected due to an accident or homicide.
In early stages of body decomposition, recognizing violent deaths is a relatively straightforward
task established on the basis of a variety of trauma differentiating evidence revealed in the course of
external or internal body examination (e.g., cutaneous injuries, skin coloring, bloodstain patterns, incisions
etc.). As soft tissues starts to decompose indicators of traumatic events lose their distinctiveness
and signature diagnostic features. Their appearance turns more generic and prone to misdiagnoses.
For heavily degraded, partially or completely skeletonized, cremated or mummified human remains,
assessment of injuries is mostly dependent on examination of hard tissues (bones, teeth). In these
cases, bones and teeth represent only direct evidence of violence, offering if not a permanent, then
a highly resilient record of a death event.
Assessment of skeletal injuries and determination of their relevance to medicolegal death investigations
is one of the main tasks a forensic anthropologist is asked to conduct in the course of a routine
examination of human skeletal remains. More importantly, due to a complex character of skeletal injuries
forensic anthropologists are increasingly being asked to consult on skeletal trauma cases, even in
instances, which are within the competence of a forensic pathologist.
An in-depth list of tasks in forensic skeletal trauma analysis includes determination and identification
of present injuries, assessment of their extent and life threatening nature, recognition of their origin
and timing, and identification of seminal agents responsible for the skeletal alterations. The occurrence
of a skeletal injury almost exclusively presents itself as a fracture. A fracture occurs when the hard
tissue (bone, tooth, calcified cartilage) cannot sustain external force. Occasionally, an infraction, bone
loss or skeletal deformity may also suggest a traumatic event. A deep and complete understanding on
how bones break and how they heal is paramount for interpreting skeletal trauma. The healing process
is one of the few direct indicators of injury timeline. Timing sets skeletal lesions into one of three main
categories. Antemortem trauma is a skeletal lesion that occurs prior to death and exhibits signs of the
healing process (e.g., blunt edges, bony callus, remodeling). Perimortem trauma is a lesion that occurs
at or around the person’s time of death and postmortem trauma is any ex vivo damage, frequently
presented in form of taphonomically or thermally induced fractures.
A key aspect in investigating traumatic events is to identify the mechanism that caused the injury. The
known mechanism may reveal behavioral preferences of a perpetrator, classification of a crime or a
sequence of events that had led to the injuries. Three main categories of skeletal trauma are generally
recognized in regards to seminal weapons – gunshot, blunt force trauma and sharp force trauma.
Blunt force trauma results from low-velocity mechanical force involving broad or blunt surfaces (e.g.,
blunt instrument, falls, energy waves or perpetrator’s body parts). Sharp force trauma occurs from a
low-velocity impact of an instrument with a cutting surface (e.g., knife, sword, machete, axe, but also
saw). Gunshot trauma results from high-velocity interactions between a bone and a moving projectile.
Appearance and patterns of skeletal injuries are relatively distinctive for each of the categories. Nevertheless,
the number of endogenous and exogenous interfering factors (deceased person’s age/sex, impact
site, force, velocity, instrument size etc.) is nearly infinite and atypical, ambiguous or transitional
characteristics may occur and eventually complicate the diagnosis. For gunshot trauma having resulted
from a single fire, skeletal remains present typically one rounded entrance wound, a larger irregular
exit wound, and radiating and concentric heaving fractures. Blunt force trauma is associated
with linear, compressed or comminuted fractures. Occasionally, the causative instrument may leave
a matching impression at the impact site. Sharp instruments produce weapon-specific injuries (stab
wound, cut mark, chop marks, saw marks), which are reflective of blade characteristics (e.g., striations,
kerf shape etc.).
Regardless of the type of trauma, the most attention has been given to injuries localized in human
skull, most likely as a reflection of their severity and life-threatening nature. However, this does not exclude
postcranial bones from being subject of skeletal injury analysis. Long bones of upper and lower
198
Chapter 4: Forensic Skeletal Traumatology
limbs are particularly frequent in assessment of dismemberment trauma. Dismemberment, the intentional
separation of body segments, is a specificity of forensic skeletal traumatology. Falling under
the umbrella of postmortem injuries, these cases are mostly associated with perpetrator’s attempts
to obliterate evidence of the cause or manner of death and/or to prevent identification of the victim.
Commonly, multiple sharp force trauma, such as saw or cut marks are evidence of dismemberment
and subjects of a thorough analysis aimed at differentiating class and individual features of a causative
tool. These cases may also include skeletal remains that were otherwise altered, for instance,
boiled, burned or submerged in acid.
Alike the areas discussed in the previous chapters, skeletal trauma analysis is an ever-developing area
of forensic anthropology whose existence is nowadays linked to advanced imaging and computation
technologies. The virtual autopsy approach capable of displaying the spatial arrangement of fractured
skeletal elements has made a great impact on diagnoses of violent events. High-resolution imaging
technologies (e.g., microCT units, 3D stereomicroscopy, scanning electron microscopy) made it possible
to examine patterns of osseous damage on a microscopic level. 3D computer modeling has been
shown helpful in trauma-induced fragmentation of skeletal remains allowing to exploit various scenarios
of a traumatic event without destroying the original evidence. And, matching 3D digital lesions
with 3D digital copies of a suspected causative instrument has become more than a rarity.
Still, novel mechanisms of force that can damage human bones rapidly increase and experts are presented
with less distinctive, even oddly looking skeletal injuries. Although for an anthropologist the
descriptive aspect of trauma analysis will likely continue to prevail, the knowledge of dynamic and biomechanical
principles of skeletal injuries are becoming more and more important to draw proper
conclusions. Experimental and computation designs are critical in understanding how and why bone
fractures occur. Advanced mathematical modeling, specifically using finite element analysis (FEA), has
become a welcome aid for the validation of the bone‘s biomechanical behavior. The principal benefit
of the computer modeling is that it allows working with human bones, albeit of virtual nature, varying
in size and shape. The drawback is that bone is an anisotropic viscoelastic material, whose complex hierarchical
structure with a wide range of intra- and inter-individual variations makes the mathematical
simulations extremely computationally expensive unless a significant degree of simplification is involved.
To this end, the majority of research remains focused on experimental modeling – ballistic and biomechanical
experiments. Conducted typically on animal cadavers and artificial bones, experimental
modeling analysis has been searching for alternative, more appropriate, model materials that would
reproduce patterns of real-life injuries more authentically. The rapidly developing technology of 3D
printing with the use of bioprinting materials combined with CT-generated digital models of victim’s
body segments is regarded as having the potential of substituting human bones in foreseeable future.
There is little doubt that these new techniques bring a great improvement in the studies of skeletal
trauma. Yet, similar to the other novelties with judicial consequences, not all of them have been proven
admissible in courts. Although computer-aided approaches in skeletal traumatology can be considered
extensions to the traditional analytical techniques, a valid argument against presenting such
evidence in courts is that the non-materialistic nature of digital data makes them prone to careless,
possibly untraceable modifications. As most of the skepticism derives from lacking standard procedure
guidelines, which would fulfil the legal implications, only further research can fill these gaps.
Two papers have been appended to this chapter, the first of them “Agenesis of the Superior Cornua of
the Thyroid Cartilage. A Rare Variant of Medicolegal Importance” is a published case report, which raises
awareness about rare anatomical variants occurring in the laryngohyoid complex. The paper falls in
line with the triple studies on hyoid bone featured in Chapter 1. Specifically, it stresses the importance
of neck elements from the diagnostic viewpoint. Human hyoid bones and cartilaginous laryngeal elements
are highly variable, yet sometimes neglected, structures whose importance rises if evidence of
trauma due to sustained neck injuries is implied. Isolated neck trauma may suggest both self-inflicted
199
Chapter 4: Forensic Skeletal Traumatology
injuries (suicide by hanging) and trauma due to perpetrator’s aggression (manual or ligature strangulation,
compression with a blunt instrument). The knowledge of anatomical variants and morphological
anomalies is imperative for a correct diagnosis at autopsy as well as in the process of anthropological
examination.
The chapter closes with the last research paper in the volume entitled “Virtual Approach to the Assessment
of Skeletal Injuries in Human Skeletal Remains of Forensic Importance”. It is a submitted manuscript
draft (submitted on August 18th
, 2016), whose objectives were drawn from three real forensic anthropology
cases involving skeletal trauma. The case study, developed in collaboration with the Forensic
Analysis Lab at North Carolina State University, outlines the first attempt to test a 3D virtual approach
in the forensic context featuring the technology of 3D printing. The study examines two cases of
gunshot wounds and dismembered skeletal remains with evidence of saw marks. One of the presented
cases also demonstrates that the physical environment in which a body is deposited can largely alter
the appearance of a skeletal injury.
201
Chapter 4: Forensic Skeletal Traumatology
4.1. Agenesis of the Superior
Cornua of the Thyroid Cartilage
A Rare Variant of Medicolegal
Importance
by Petr Hejna, Martin Janík,
Petra Urbanová
published in American Journal of Forensic Medicine and
Pathology 36/1 (2015)
207
Chapter 4: Forensic Skeletal Traumatology
4.2. Virtual Approach to the
Assessment of Skeletal Injuries in
Human Skeletal Remains of
Forensic Importance
by Petra Urbanová, Ann H. Ross, Mikoláš Jurda,
Ivana Šplíchalová
submitted to Journal of Forensic and Legal Medicine,
August 18th
, 2016
Petra Urbanová
New Dimensions of
Forensic Anthropology
NewDimensionsofForensicAnthropology