Introduction
While the process of identification of human remains incorporates many different factors, the face is regarded as a unique feature that can be used to rightfully identify a person. The face plays an important role in identification when DNA or fingerprints examinations cannot be used. Porter and Doran (2000) observes that while fingerprints are more reliable sources of identification, photographs of the face help the layperson to make identification by comparing the suspect in question with his/her photograph or sketch. The aim of forensic facial reconstruction is to use a skull to build a face that can be used to recognize the person to whom the skull belongs. A successful forensic facial reconstruction will result in a face that can be easily identified accurately as that of the person to whom the skull belongs. The significance of facial reconstruction techniques in forensic work have resulted in the development of a number of facial reconstruction methods. This paper will explain the basis of forensic facial reconstruction and critically assess its usefulness. A review of some of the most popular methods of facial reconstruction will be made and their inherent merits and demerits highlighted.
Basis of Forensic Facial Reconstruction
Facial reconstruction is the process of constructing faces (normally of unidentified persons) from skulls. This process is based on the hypothesis that “the anatomy of the skull is intimately associated with the surface anatomy of the face” (Oxenham, 2008, p.133). However, this relationship is not easily determined since the soft tissues of the body are not uniformly placed over bones. For this reason, the accuracy of facial reconstruction methods is controversial and some researchers express doubt in the ability of the methods to reproduce a face that can be easily recognized correctly as the target individual. In addition to this, the skull does not contain any individual characteristics such as birthmarks, moles, or scars which can be used to identify a person.
Clement and Murray (2005) acknowledge that the task of completely reconstructing a face from a bare skull is extraordinarily difficult since few specific relationships between hard and soft tissues are known. The relationship between the skull and the face is complicated by the presence of fat, muscle, and connective tissue. Stephan and Simpson (2008) reveal that discriminative information is not provided about any single anatomic component of the face while determining facial soft tissue depths. In addition to this, facial approximation techniques are not able to determine features such as face colors and textures from the skull. Therefore, it is still impossible to determine the skin depth with 100% accuracy from the bare skull even with the large quantity of data from a great cross section of the human population.
In spite of the difficulty in achieving accurate anatomical construction of the face, forensic facial construction has been successful in generating leads or preliminary identifications in many forensic cases. Wilkinson (2004) reveals that police use forensic facial reconstruction when they do not have a suspect for identification. The face is considered the most potent part of the body to aid in such identification efforts. Oxenham (2008) declares that the face is the bodily feature with almost unique configurations and this gives the facial reconstruction methods a great identification value.
In addition to helping identify whom the skull belonged to, facial reconstruction is also an important tool in excluding suspected individuals (Damas, 2011). Forensic facial reconstruction is also used by Anthropologists to recreate the facial images of early man and hominids. Such an application helps anthropologists in their studies of the evolutionary changes between hominids and modern man. Technological advances over the last few decades have enabled anthropologies to employ a number of techniques that provided for more accurate identification.
Facial Reconstruction Techniques
The early attempts at facial reconstruction were aimed at matching skulls of famous people with their portraits and sculptures. The unidentified skulls were superimposed on the images of the suspected owner to see if they matched. However, the significance of facial reconstruction in recreating faces of unidentified individuals became greater and there was a move in research to come up with techniques to recreate the faces of unknown individuals from their skulls (Wilkinson, 2004). While facial forensic reconstruction primarily relies on the skull to come up with an approximation of the face, physical remains on the skeleton are useful in the process. Remains of soft tissue assist the practitioner to better approximate soft tissue thickness and this eliminates the errors that are unavoidable in the techniques employed to try to estimate the soft tissue thickness.
2D Reconstruction
An outline drawing of skulls in frontal and lateral views is produced by sticking tissue depth markers on the skull that is to be identified. Oxenham (2008) demonstrates that major sutures of the skulls (also known as anthropological landmarks) are focused on when sticking these depth markers. From the photographic prints obtained through this method, artists are able to produce 2-D facial reconstructions. Wilkinson (2004) reveals that using this method, police artists have been able to reconstruct the face of decomposed bodies and correctly identify the original owner of the skull.
The traditional 2-D facial reconstruction was done manually with drawings of the face produced on overlays superimposed onto images of the skull (Rosalie, 2008). In the manual approach, the practitioner might draw the facial musculature on the initial overlay or dray the finished face directly over the skull images. The 2D technique was improved by Caldwell who proposed the use of life-size frontal and lateral views of the skull and then adding tissue depth data to the same using Krogman’s guidelines.
The traditional 2-D reconstruction method has achieved significant advancements in recent decades due to the application of computer programs in the process. Computerized 2D facial reconstruction involves the use of computer software designed specifically for this purpose. The software program is used to create a facial composite from an image of the skull. Facial outlines and features are then added from a database maintained by the program. These facial features are dependent upon the skeletal structure and its origins.
The renowned Russian Anthropologist Mikhail Gerasimov suggested that major problems in accuracy could occur when the investigator fails to pay enough attention to muscle structure and the interdependence of the form of the face and the peculiarities of the skull (Wilkinson, 2004). As such, the accuracy level of 2D reconstruction depends on the experience of the practitioner. Ullrich and Stephan (2011) reveal that Gerasimov’s facial approximation technique, which considered muscle structure, achieved close to 100% accuracy.
Strengths
When computer software programs are used in the 2D reconstruction, the process is speeded up and an approximation can be produced in little time. Variations to the final image can also be implemented with relative ease using the editing features of the software. This method does not require the use of sophisticated machinery to come up with the facial approximation. It can therefore be used by laboratories that lack modern apparatus. This factor is of great importance since a lack of resources sometimes acts as a hindrance to forensic teams carrying out facial reconstruction.
Weaknesses
This facial reconstruction technique is also prone to the individual bias of the reconstructor. Wilkinson (2004) observes that the perceived subjectivity introduced by the reconstructor will greatly influence the outcome of the process. Accuracy levels in 2D reconstruction using the manual level are highly dependent on the abilities of the artist. The level of skill and knowledge of the individual reconstructor is the main determinant of the accuracy of the reconstruction. The reconstruction practitioner utilizing this method must be experienced enough and he/she must possess a thorough understanding of facial anatomy and physiology in order to come up with an accurate enough reconstruction of the face.
3D Sculpture
The 3-D method was developed by Gatliff Betty from the work of Krogman. Gatliff discovered the importance of facial asymmetry in her work reconstructing the face of a Native American man. Wilkinson (2004) documents that Gatliff reconstructed only half of the face in the case and created a full face by using mirror image photography. There are two 3-D facial reconstruction techniques; the American and the Russian methods. In the American method, the skull is mounted on a stand and an average tissue thickness data from a variety of tables relating to different ages, ethnic groups and sexes is used. The practitioner selects the most appropriate set of tissue thickness data and cylinders of vinyl eraser are cut to the appropriate thickness. These pieces are glued to the skull at the appropriate anatomical points and modeling clay is used to connect these markers by strips. By doing this, a rough contour map of the surface of the face is created and the remaining open faces are filled to create the semblance of a face. An artist is then used to sculpture the facial features such as the eyes, lips, nose, and ears. Wilkinson (2004) observes that the facial features are based on details obtained from the skull which might point to the origins of the skull.
Strengths
The availability of a 3D model of the face results in a more realistic outcome and this helps in the identification process. Stephan and Henneberg (2001) document that the ability of identification is significantly higher than chance when a 3D model is used. Accuracy levels can be significantly high especially if the Russian method which requires the construction of the facial muscles is used (Ullrich & Stephan, 2011). 3D sculptures allow the practitioner to feel the surface of the skull during analysis and this makes it possible for important skeletal details to be added during the facial reconstruction process. Appropriate Facial features are very useful in helping lay people make accurate identifications.
Weaknesses
Only a few dozen points with height information are used for reconstructions, the largest parts of the face are build upon interpolated information, and this decreases the inaccuracy of the reconstructed face. If the inaccuracy level is too high, it will be impossible to make a positive identification using the reconstructed face. Another problem with this technique is that it is time consuming and the practitioner requires deep anatomical and forensic knowledge in order to come up with an accurate reconstruction.
3D CT scan
CT scans are very useful when it is impossible to use the original skull of the unknown person. A copy of the individual’s skull can be manufactured with precision using CT scans in a process known as stereolithography. The CT scan of the skull can be printed out with a 3D printer or it can be manipulated as a model in the computer. Manipulation of the model enables the process of tissue reconstruction to be completely computerized. The dowels which act as an orientation for the process of reconstruction are directly integrated into the CT image through software and this makes editing very easy.
Strengths
The 3D scan yielded by CT is free of any geometrical distortions. CT scans therefore serve as good starting points to computer aided craniofacial reconstruction. Choras (2011) affirms that the resulting scan is very accurate and from it, equipotential planes can be identified and triangulated. The CT scan results in distortion free 3D images of the skull since it is a non-invasive process. Choras (2011) documents that it serves as a point to start from where several ways of tissue reconstruction can be implemented.
The practitioner can manipulate the height of the dowels and therefore change the appearance of the face with ease. Clement and Murray (2005) state that this ease of manipulation makes the facial recreation process fast. CT scan images can be saved as data files which will be used later to identify the person even in the absence of the original skull. Such a consideration is very important in culture where human remains are destroyed through cremation.
Weaknesses
This method requires specialized equipment to produce the 3D scans. This equipment might not be available in all laboratories which makes this method inaccessible to some forensic practitioners. When reconstructing a face from the obtained scan, averaging methods are used to compute the tissue size. The heavy reliance upon averaging methods in this technique increases the rate of error in the final reconstruction.
Superimposition
Superimposition is a technique that involves layering the skull of the subject with the image of an individual suspected to be the owner of the skull. Ishii et al (2011) reveal that PC-assisted craniofacial superimposition is the most frequently used method. A photograph of the suspected skull owner is superimposed on an x-ray image of the skull to see if they match. For this method to be used, all residual soft tissue has to be removed from the skull. This will increase the accuracy of the superimposition technique. In instances where removing the soft tissue is not desirable, CT images can be used in superimposition since the CT scan image is almost completely consistent with real skulls.
Strengths
Superimposition is very effective in eliminating suspects since it can accurately determine that the skull does not belong to a particular person. The accuracy level of this method is very high especially if good quality photographs of the suspect are available. This method can assist in the identification of the remains of a person even after the physical skeleton has been disposed off. Ishii (2011) elaborate that by saving the skull images as digital data, superimposition can be applied to the facial images of suspected persons at later points in time.
Weaknesses
A major disadvantage of this method is that the forensic team must have a list of suspects to match against the skull. As such, some knowledge about the identity of the skull is required for this process to be used. Without such knowledge, it is impossible to use this method in facial reconstruction efforts.
Lack of good quality pictures impedes on the accuracy level of superimposition since clear images are desirable in the process. Aulsebrook et al. (1995) notes that most people smile when having their photos taken which makes it hard to superimpose the mouth. The lighting in most photographs is also not clear with some parts of the face obscured.
Conclusion
This paper set out to discuss forensic facial reconstruction and discuss some of the common methods used and their merits and demerits. It began by highlighting the importance of the face in identifying an individual. The major facial reconstruction techniques: 2D, 3D sculptures, 3D CT scan, and Superimposition have been reviewed. The paper has underscored how facial reconstruction software has been used to overcome some of the weaknesses of the manual methods. The paper began by noting that facial reconstructions are not identical to the picture image of the individual. All the forensic facial reconstruction methods techniques employed have some weaknesses and it is impossible to eliminate all the problems and come up with an exact representation of the subjects face. Even so, the techniques are able to produce a face that is sufficiently similar to help in the identification of the person by someone who knew the skull owner well.
References
Aulsebrook, W.A., Iscan, M.Y. Slabbert, J.H., & Becker, P. (1995). Superimposition and reconstruction in forensic facial identification: a survey. Forensic Science International, 75 (1), 101 – 120.
Choras, S.R. (2011). Image Processing & Communications Challenges. Boston: Springer.
Clement, J.G. & Murray, K.M. (2005). Computer-Graphic Facial Reconstruction. NY: Academic Press.
Damas, S., et al. (2011). Forensic identi¯cation by computer-aided craniofacial superimposition: a survey. ACM Computing Surveys, 43(4), 1-31.
Ishii, M. et al. (2011). Application of Superimposition-Based Personal Identification Using Skull Computed Tomography Images. J Forensic Sci, 56 (4), 960-966.
Oxenham, M. (2008). Forensic Approaches to Death, Disaster and Abuse. Sydney: Australian Academic Press.
Porter, G., & Doran, G. (2000). An anatomical and photographic technique for forensic facial identification. Forensic Science International, 114 (1), 97-105.
Rosalie, D. (2008). Egyptian Mummies and Modern Science. Cambridge: Cambridge University Press.
Stephan, C.N., & Henneberg, M. (2001). Building faces from dry skulls: are they recognized above chance rates? J Forensic Sci, 46(3), 432–440.
Stephan, C.N. & Simpson, E.K. (2008). Facial Soft Tissue Depths in Craniofacial Identification (Part I): An Analytical Review of the Published Adult Data. J Forensic Sci, 53 (6),1273-1279.
Ullrich, H., & Stephan, C. (2011). On Gerasimov’s Plastic Facial Reconstruction Technique: New Insights to Facilitate Repeatability. J Forensic Sci, 56 (2), 470-474.
Wilkinson, C. (2004). Forensic Facial Reconstruction. Cambridge: Cambridge University Press.