SUBJECT

Paper No. and Title PAPER No.11:

Module No. and Title MODULE No.24: Facial Superimposition-I

Module Tag FSC_P11_M24

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

TABLE OF CONTENTS

1. Learning Outcomes

2. Introduction

3. Two-Dimensional Images

4. Three-Dimensional Skull Models from Limited Two-Dimensional Data

5. Three-Dimensional Skull Models from Clinical Imaging

6. Craniofacial Superimposition

7. Video Superimposition

8. Summary

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

1. Learning Outcomes

After studying this module, you shall be able to-

 Know about the importance of facial superimposition  Learn about the various techniques of facial superimposition

2. Introduction

In some forensic investigations the usual methods utilized for human identification can be unsuccessful, and the police may have few clues as to the identity of an individual. The majority of identification techniques require a known individual with whom to compare data, such as DNA, , or dental records, and where there are no suspects for identification, it is practically impossible to compare data with records from an entire population. In these circumstances the police may employ less definitive methods in an attempt to focus on a population from which the individual may be identified. Facial reconstruction is one of the methods that are frequently employed in such investigations.

Facial reconstruction (otherwise known as facial approximation) is the process utilized to reproduce the facial appearance of an individual and includes a number of different procedures. Traditionally, facial reconstruction has involved the analysis of skeletal detail to determine facial morphology. However, facial soft tissues may be present as partially decomposed, damaged, distorted, or preserved remains. In these cases it may not be appropriate to present images of the remains to the public, and facial reconstruction is utilized to visualize facial appearance. In these circumstances the practitioner will analyse the soft tissues rather than skeletal detail, although it may be possible to analyse both soft and hard tissues where clinical imaging or dissection are employed.

Different approaches in this field have created some confusion with regard to the reliability and presentation of a facial reconstruction within a forensic scenario. Some practitioners attempt to reproduce a facial “type,” an approximation of the facial proportions and morphology that relies on sets of average tissue data and facial templates relating to the sex, age, and ethnic group of the individual (Evison et al. 1999, Vanezis et al. 2000, Stephan 2004).

Facial superimposition is one of the techniques that come under forensic facial reconstruction. It is the technique by which a particular skull recovered can be matched with the photograph. It is very helpful for identification. More recent photograph is better. Even photograph of lateral view can be used. Two negatives are prepared one of the person and the other of the skull.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

Appropriate magnification is done to achieve maximum alignment and where inter-pupillary distance of both exactly matches. It is a known fact that the inter-pupillary distance of a person never changes regardless of his age. The two negatives are then superimposed by keeping over each other and various points are compared. The canthus, the nasion, the nasal spine and lower border of nose and upper jaw, the supraorbital ridges, angle of jaw, external auditory meatus and the teeth etc. are compared and a photograph is taken which is analysed for similarities and dissimilarity to arrive at a conclusion. This test has more of negative value as it can definitely rule out a combination of a photo and a skull though only a possibility of a match can be established.

3. Two-Dimensional Images

In certain circumstances, photographs or x-rays of the skull may be the only available information, and access to the original specimen may be restricted due to legal matters, inaccessibility, destruction, or damage. Where images of the skull are considered adequate, two-dimensional facial reconstruction can be performed. Ideally, the skull should be in the Frankfurt Horizontal Plane (FHP), and for 35mm photography, a lens with a focal length of 100–200mm and a camera position of no less than 12 ft/3 m from the skull will avoid distortion. When there is access to the original specimen, the facial anthropologist should take scaling measurements or include a visual scale in the foreground of the images. If possible, tissue depth markers should be attached to the skull prior to production of the images to provide the maximum number of tissue indications. When this is not possible, but a scale has been included, some of the tissue depth markers can be indicated on overlays superimposed onto the images of the skull.

4. Three-Dimensional Skull Models from Limited Two-Dimensional Data

When access to the original specimen is not possible, either directly or via three-dimensional clinical imaging, methods of three-dimensional model production from two-dimensional data (such as radiographs, photographs, and craniometrics) may be utilized. Radiographs and/or photographs are used as templates, and multiple views are aligned using cranial points as registration marks. Computer modeling software is then employed to create a three- dimensional model (Wilkinson 2005) or distort a template mesh (Davy et al. 2005) to reproduce the skull morphology. Extrapolation of surface morphology between the views is inevitable, and the more views that are available, the more accurate the resulting three- dimensional model of the skull will be.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

As these methods include a certain degree of estimation and loss of detail on the surface of the bone, any resulting facial reconstruction or approximation would require photographic records of the skull and an appreciation of the decrease in accuracy of any resulting face. Care must be taken when establishing anatomical landmarks.

5. Three-Dimensional Skull Models from Clinical Imaging

Where soft tissues are present, access to the original specimen is restricted or computer-based facial reconstruction or approximation is planned. Cross-sectional data created by x-ray computed tomography (CT) or surface scans can be employed to produce a three dimensional digital model of the skull (Spoor et al. 2000). Sometimes replica skulls may be produced from digital data using stereo-lithography (Hjalgrim et al. 1995) or another form of three- dimensional model manufacture (Seitz et al. 2005). Where digital data from surface scans and CT scans are employed the limitations of the data must be taken into account. With CT data, the slice thickness, scan plane, spatial resolution, filters, and angle of rotation will all affect the resulting three-dimensional model. In addition, dental filling and appliances will cause artefacts and may require some manual intervention.

A variety of software packages render the surface cross-sectional CT data by extracting selected tissues and visualizing the tissues as three-dimensional models for use in computer- based facial reconstruction or approximation systems, or for production of a physical replica. Surface rendering involves segmentation (isolation of the tissue by thresholding), interpolation between the slices to create a smooth surface, and illumination of the surface. Spoor and his colleagues (2000) state that “However, improved visual representation [is, this] does not imply that the image is more accurate as well. The extent to which the reconstruction reflects reality primarily depends upon limitations inherent to CT or MRI.”

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

Fig.1: Common facial anthropometry points

With surface scan data morphology, details may also be lost due to discrepancies at apertures, fossae, and holes where the scanning mechanism cannot visualize the surface. In addition, hair on the face and the head may cause surface artefacts. As surface scanners produce a series of profiles from a number of viewpoints, which then require computational reconstruction of the object and subsequent manual editing to ascertain clean, raw data sets, there may be inconsistencies to the original specimen. Where possible, attention should be given to the original specimen, so that errors are not perpetuated to the facial reconstruction.

There are many surface scanners currently available, including laser (Kau et al. 2005, Moss et al. 1989, Bush and Antonyshyn 1996), photographic projection (Yamada et al. 1999, Motoyoshi et al. 1992, Techalertpaisarn and Kuroda 1998, Curry et al. 2001, Siebert and Marshall 2000), or holographic (Bongartz et al. 2000, Giel et al. 2004), and three-dimensional models will be created that can be imported into computer-based facial reconstruction or approximation systems or employed to produce a three-dimensional replica.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

Where the skull is fragmented and each piece has been scanned, reassembly can take place using some computer-based systems. There are computer-based systems (Wilkinson 2003, Sanghera et al. 2002, Bibb et al. 2000) that allow manual reassembly on the computer in three dimensions. Computerized skull reassembly is much more efficient and rapid as no support mechanisms are necessary. Computerized re-modeling of missing fragments is also easier and less time-consuming with the computer-based systems and perhaps involve a few hours of work rather than days or weeks. Some fragment-edge detail may be lost on the digital models, but access to the original specimens will avoid any resolution problems.

When replica skulls are produced from digital models using stereo-lithography (Hjalgrim et al. 1995) or three-dimensional printing (Seitz et al. 2005), the models will suffer from the same limitations as the original digital data. In addition, problems relating to the replication procedure may also occur, such as material fragility (susceptibility to water damage or crumbling), material resistance (difficulties with drilling or inability to bond with glues), or smoothing of some of the detail to smaller bones.

6. Craniofacial Superimposition

Photographic superimposition in its simplest form includes creating a photographic image of the skull that can be superimposed on an ante-mortem photo of the person. Superimposition assumes that photographs accurately reflect the details of the face. Although using a camera solves some of the problems of direct cranial reconstruction, such as tissue thickness or the ability of the artist, new problems are created. One of the most critical is how to photograph the skull in the exact position and at the same distance as the ante-mortem photo.

Within craniofacial superimposition, a number of different techniques have been proposed by investigators not only in the United States and Canada but also in Europe, China, and Australia.

One of the most famous examples of early attempts at craniofacial superimposition was the Buck Ruxton case in Scotland in 1935. Dr. Ruxton had done away with both his wife and female housekeeper. Because of his medical background, he knew that various parts of the body, such as fingerprints and ears, could be used to identify the body and thereby lead to the killer. Therefore, he dismembered both victims and further dissected soft tissue from bone. He then packaged the remains and deposited them in parcels across the landscape. As more bundles were found, the difficulty of identification superseded shock over this grisly case. It became clear that two women had been the victims and that they were roughly the same age.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

Eventually, suspicion focused on Dr. Ruxton whose wife and housekeeper were mysteriously missing. Yet, there was no way to identify the dismembered remains as these two women. With good ante-mortem photos of each victim available, John Glaister, Professor of Forensic Medicine at the University of , decided to try to identify both women through photo superimposition.

One of the principal difficulties in this technique is placing the skull at the same distance and in the position as the head in the picture. Fortunately for the case, Mrs. Ruxton had been photographed shortly before her death. The local town photographer was able to reconstruct the distances and angles in the ante-mortem photo by using the same gown and tiara in the post-mortem photo. The match between ante and post-mortem photos along with other evidence was good enough to identify Mrs. Ruxton and to convict Dr. Ruxton.

In more recent times, others have experimented with this technique. One of the most notable techniques was developed by Tadao Furue, forensic anthropologist at the U.S. Army Central Identification Laboratory (CIL), first in Japan and then in Hawaii. Furue used a system that included a large format camera, front surface mirror, and a beam splitter or partial mirror to create his superimpositions. This technique allowed him to use an original life photo to position the skull correctly. In this way, a transparency photo of the skull could be laid over the life photo to see if they matched.

N.S. Klonaris and Furue used a variation of this technique to match a fragment of maxilla without teeth with an ante-mortem dental x-ray. In fact, forensic odontologists probably use this technique more than other specialists because the structure and position of teeth, the surrounding bone of the jaws, and odontological repairs provide a great variety of structure that can be compared between ante and post-mortem images.

7. Video Superimposition

The introduction of video cameras and computers has taken photo superimposition a step farther. Instead of using mirrors and still cameras, two video cameras are used: one focuses on the skull, the other image is centered on the ante-mortem photo. As with still photos, the ante-mortem image is used to orient the skull. The difference is that a video mixer is used to superimpose the two images through the cameras. Regardless of the specific techniques of facial reconstruction or craniofacial superimposition, the question remains, “How similar do the skull and photo have to be in order to match?” Most researchers agree that craniofacial superimposition is a good technique for excluding potential matches.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I

If the ante and post-mortem images do not fit, they probably represent two different people. However, is the technique so accurate that if two images match, they represent one and only one person? Unfortunately, a detailed evaluation and test of the various techniques has not been done systematically, so this question remains unanswered. There are cases in which each of these techniques has been used successfully; yet, both can be misleading. Virtually all the investigators who have proposed one of these techniques say that it should never be used by itself to establish identity; other evidence is always required. Facial reconstruction or craniofacial superimposition may prove to be useful, but it should not be relied upon by itself to determine identity.

8. Summary

 Facial reconstruction (otherwise known as facial approximation) is the process utilized to reproduce the facial appearance of an individual and includes a number of different procedures.  Different approaches in this field have created some confusion with regard to the reliability and presentation of a facial reconstruction within a forensic scenario.  Facial superimposition is one of the techniques that come under forensic facial reconstruction.  The canthus, the nasion, the nasal spine and lower border of nose and upper jaw, the supraorbital ridges, angle of jaw, external auditory meatus and the teeth etc. are compared and a photograph is taken which is analysed for similarities and dissimilarity to arrive at a conclusion.  When access to the original specimen is not possible, either directly or via three- dimensional clinical imaging, methods of three-dimensional model production from two-dimensional data (such as radiographs, photographs, and craniometrics) may be utilized.  With CT data, the slice thickness, scan plane, spatial resolution, filters, and angle of rotation will all affect the resulting three-dimensional model.  Photographic superimposition in its simplest form includes creating a photographic image of the skull that can be superimposed on an ante-mortem photo of the person.  One of the most famous examples of early attempts at craniofacial superimposition was the Buck Ruxton case in Scotland in 1935.  Facial reconstruction or craniofacial superimposition may prove to be useful, but it should not be relied upon by itself to determine identity.

FORENSIC SCIENCE PAPER NO.11 : Forensic Anthropology MODULE NO.24: Facial Superimposition-I