Introduction

One of the most helpful features in the identification of a human is their face and when the soft tissue becomes damaged or even missing, what we have left for identification are the bones of the skull. Historically, forensic anthropologists have used estimations of facial soft tissue thickness on certain parts of the skull to reconstruct a face based on previous data collected from human cadavers of both sexes. Whenever these estimations were applied, artistic adjustments were made to create a final facial form. Usually, this is done with clay representations of the face applied to the skull. This process has been and still is tedious, often ending in only one face based on the sex determined by forensic specialists. Yes, but, what if the sex estimation was wrong? If a forensic anthropologist was not involved in the skeletal analysis (especially for cases decades old), sex may not have been correctly determined.

Current averages for determining this have also changed due to an increase in data and new methods. Even if a forensic anthropologist was involved, sex may be difficult to estimate from the skeletal remains, presenting a problem for attempts to generate a facial approximation.

If a face was reconstructed around a female skull using clay, attempts at a positive identification on the skull may not be achieved if the face was designed for the male sex. The process to correct the reconstruction would take a lot of time and work depending on the medium. Thankfully, with the abilities of new statistically based facial approximation software, we can avoid tedious clay models when plotting on a skull to reconstruct the face. By simply selecting a different reference group option (sex) from the software database, we can quickly generate a different face with the hopes of achieving a positive identification faster and easier than doing such work with physical clay. Through usage of different software, we have compiled sets of data to support that the main differences between male and female face reconstruction samples exist in the medial orbital, outer lip edge, and superior ear regions of the face. Though subtle, these main differences show the existence of average sexual dimorphism data within facial reconstruction.

Materials

We were given a set of skulls from both sexes ranging in sexual characteristics. The skulls given were QIN-HEADNECK-01-0109 (male), QINHN01_0041_6624 (female), QINHN01_0112_Skull200

(female), QINHN01_0003_Skull200 (male), QINHN01_0224_2223_skull200 (female), and

QINHN01_107_Skull200 (male). All the skulls are from the Texas State University catalogues and were provided by Professor Terrie Simmonds-Erdherdt. The skulls ranged in sexes between male and female with variances in sex categorization. By that we mean some known to be female skulls and male skulls had sex characteristics belonging to the opposite sex. The skulls were digitally scanned and presented as

.stl files.

For the facial reconstruction software, we used the treatment and Increased Vision for

Medical Imaging (TIVMI) software. Created by Bruno Dutailly, “(TIVMI) is a software developed in the PACEA laboratory, UMR 5199 of CNRS, Université Bordeaux 1, and ministère de la culture. It is designed for anthropology researchers for precise and reproducible measurements in 3D and medical imaging when other software do not respond to their needs^[1].” This software was specifically used for facial reconstruction using the digital skulls as a reference point for the skulls. Using previously collected soft

tissue averages, 78 landmarks were placed along the skull with each point denoting a skin depth.

The other software used was Meshlab. Meshab, according to the website, “is an open source, portable, and extensible system for the processing and editing of unstructured 3D triangular meshes. The system is aimed to help the processing of the typical not-so-small unstructured models arising in 3D scanning, providing a set of tools for editing, cleaning, healing, inspecting, rendering and converting this kind of meshes^[2].”

Methods

With each skull we followed a series of steps to reconstruct the face. We started with using the TIVMI software to load a skull as a mesh and then transferred the file into a AFA3D file. Once transferred, we placed 78 landmarks onto the skull by a specific anatomical order as pre-ordained by the software. Once the landmarks were placed, we saved the XYZ coordinates into an excel file for our records and to compare the results when we switched the sexes. After saving to the file, we reconstructed the face using the AFA3D facial reconstruction option.

This is where we ran into some issues with the software. When reconstructed, the facial skin averages went astray. This caused a skewing of the face like a piece of gum when stuck on both a shoe and pavement and the resulting shape after the two-part ways. After seeing the dysfunction, we did not calculate these off-shoot skin points.

To overcome this issue, we used a different method involving the Point 3D module of the TIVMI software. We followed the same methods for AFA3D with each 3D point on the skull being placed and re-written. We then saved the points, extracted those saved points, and input those points into the AFA3D program. The major drawback in using this system was re-organizing the XYZ coordinates to fit the paradigm used for AFA3D for the sake of compatibility. Using this method, we were able to have a

fully functional facial reconstruction relative to the sex of the skull. We then saved the face XYZ points and repeated this process while switching the variables from one sex to the other sex relative for the

skull.

Using the Meshlab software, we loaded up the skull related to taken and saved facial reconstructive points and placed both the male and female faces. Turning on the filter for “Color Vertex Quality” we super positioned the male face and female face on the skull at the same time, coloring the male face only. The reason for coloring one face was to show the areas of protrusion when the male and female face were viewed at the same time. The protrusions would show where exactly the measurements differ on average due to sexual dimorphism as concluded by the average taken skull points.

Background

Throughout the world, animal species have interspecies characteristics that differ themselves between males and females (should the species be specifically di-sexual). In the human race there are, chromosomally speaking, 5 sexes; Male(XY), Female(XX), Kleinfelter’s Syndrome (XXY), Supermale(XYY), and Turner’s Syndrome(XO). Because we cannot tell biological sex chromosomally by just plain sight, males and females are the heteronormative United States and other societies norms. In forensic science, sexual characteristics are calculated through various features, most of which are from looking at the dimensions of the skull. Using a point system, lower point 1 denoting male and higher denoting point 5 female, we estimate sex by collecting and cataloging these differing dimensions. Mental eminence, brow protrusion, and Zygomatic-temporal suture position are all factored in for a rough estimate for sex.

Because we can also not view a skull at first glance in a living person, we often come up with societal patterns to induce sexual identity within gendered fashions that range from lax to strict depending on the culture. Hair length, vocal choice, and dress may be considered masculine and feminine in the culture.

Biologically speaking, the average male face is found to be larger than the average female face. Even more so, the male face develops faster than the female. According to a study done by Xanth D.G. Mallet et al, “Male faces in the (test) sample complete the development to the adult form in the younger age group and that further changes as a result of aging are relatively subtle, whereas female faces retain a youthful morphology in adulthood^[3].” So here we find that male faces differ from an average formation age standpoint compared to women. A second study done by M. De Menezes et al found that the age of the participants involved in the study allowed for differences in size and formation. The main differences found average age-related reductions of 2 mm of the upper and 7 mm of the lower lips. “Overall, age related reductions in the lower lip were larger in men (6mm) than in women (4.8mm).” De Menezes et al also found that, “With ageing, in both sexes the lower lip became prevalent over the upper lip, a modification particularly evident in women…the main problem of ageing in the upper lip is not a simple volume reduction, but a rearrangement of all labial upper and lower features^[4].”

Comparative sex studies infer specific areas of difference around the eyes and mouth area which correlate with our study results. Labial and total volumes were taken in a study done by Chiarella Sforza et al. which found, “all labial volumes and total lip area were larger in men than in women. A significant sexual dimorphism was found also for mouth and philtrum widths, and total lip height, as well as for the vermillion height to mouth width ratio.^[5]” Sandra Anic-Milosevic et al performed a similar study with lower facial morphology. Their study found, “ the lower lip in relation to the lower third of the face

height was significantly greater in females 41.20% than in males 39.70%. Farkas et al. (1984) found that lower lip height, on average occupied a similar proportion of the lower third of the face in both genders (males 38.7%; females 37.4%)^[6].” These differences in sexual dimorphism are incredibly small in percentage however it is these differences that we are calculating. The difference in our study compared the previous examples is a rare opportunity to use the same base for the same test but shoot for different variables in our equations. Rarely in nature does the same skull reproduce itself with two genders in each incarnation.

One mid facial representation of sexual dimorphism is the piriform aperture, the pear-shaped nasal opening of the skull. A study by Eric Moreddu et al found that nasal apertures in men, on average, are larger than the women’s PA. The differences for the width in the study was 1.32 mm and the length difference is 3.81 mm in favor the males. With this significant sexual dimorphism in this study, we kept the differences in mind when collecting the data.

This study was meant to study facial differences between males and females regardless of

societal aspects. It was meant to find sex characteristics at quite literal face value. In previous studies, data has been collected using dimensions for facial plastic surgery. Orthodontic and orthognathic surgery analysis is meant to repair facial skin around the lower portions of the face. Precision and balance is important for these patients involved in the study done by Bayome et al. Their study found differences in the sexes but also similarities in certain areas of the skulls in the sagittal dimensions^[7].

As our study was not interested in cross racial comparison in terms of sexual dimorphism, other studies comparing racial facial morphology have found differences in males and females of different

races. Races, by this study, are examples of populations of similar historical cultural backgrounds inferred by geopolitical land origin. By no means do we infer races as biologically different humans by such a difference that means a hierarchical categorization. The study done by J Wirthlin et al compared facial morphologies between adult Chinese and Houstonian Caucasian populations and found, “differences between gender-specific groups: 2.83 mm for the comparison of Chinese and Houstonian male subgroups, and 2.73 mm for the comparison of Chinese and Houstonian Female subgroup^[8][9].” The main differences were interesting prominences in different areas determined by the race and sex. “Caucasian males showed more prominence in the pogonion, nasal tip, and supra-orbital region. The females exhibited the same differences as the males stated above, and the Chinese females had a more depressed philtrum area than the Caucasian females^[10].”

Data Results

Of all the plotted points and calculated skin averages, we selected the 15 points that were of highlighted difference in the excel chart. In relation to areas of the face; these points made up the corner of the mouth, lower midline point of the mouth, upper midline of the mouth, center midline of the mouth, outer curvature of the nose, tip of the nose, base of the nose, inner corner of eyelid, outer corner of the eyelid, and the most anterior point of the eyeball. When calculating the data, we kept the listed anatomical order in the graphs. The order in the anatomical sense was right cheilion (1), left chelion (2), labiale inferius (3), labiale superius (4), stomion (5), right alagenion (6), left alagenion (7), pronasale (8), subnasale (9), right endocanthion (10), left endocanthion (11), right exocanthion (12), left exoconthion (13), right ocular anterior (14), and left ocular anterior (15).

It was in these areas where we noted differences in the data when we changed the sex characteristic of the facial reconstruction. In the case of skull QINHN_0112_Skull200, a biological female, the deviation for the 15 points was 0.9828 for the right chelion, 1.1414 for the left chelion, 0.4618 for the labiale inferius, 0.49 for the labiale superius, 0.7001 for the stomion, 2.3571 for the right alagenion, 1.8913 for the left alagenion, 0.6427 for the pronasale, 1.2184 for the subnasale, 0.4473 for the right endocanthion, 0.7119 for the left endocanthion, 0.5449 for the right exocanthion, 0.9889 for the left exocanthion, 0.5221 for the right ocular anterior, and 0.5767 for the left ocular anterior which are laid out in graph 1.3. All of the points in the skulls are laid out in graphs 1.1 and 1.2 below.

Skull QINHN01_0112_Skull200 (female)

1.1 Male 1.2 Female

1.3 3D Difference

The most significant difference exists in the right and left alagenion (outer curvature of the nose) and the right and left cheilion (corner of the mouth). Both of these finds correlate with the Sforza et al, Anic-Milosevic et al, and Moreddu et al studies. The photos correlating these finds show these areas as major evidence of protrusion (photo 1.1). One non-symmetrical find was a major protrusion made by the left exocanthion (left outer corner of the eyeball) which shows a graphical rise much more superior than its right counterpart in graph 1.3 (above) and can be compared to in photos 1.2 and 1.3 below.

In the case for skull QINHN)1_0003_skull200, a biological male, we find differences for 3D points

with 1.0464 for the right cheilion, 1.3273 for the left cheilion, 0.4128 for the labiale inferius, 0.4727 for the labiale superius, 0.7111 for the stomion, 2.3115 for the right alagenion, 1.9084 for the left alagenion, 0.709 for the pronasale, 1.3309 for the subnasale, 0.4544 for the right endocanthion, 0.7291 for the left endocanthion, 0.5209 for the right exocanthion, 0.9527 for the left exocanthion, 0.5507 for the right ocular anterior, and 0.5803 for the left ocular anterior. All of the male/female points are laid out in graphs 2.1 and 2.2 below.

Skull QINHN01_0003_skull200 (male)

(2.1) Male (2.2) Female

(2.3) 3D Distance

Graph 2.3 above shows the same pattern in 1.3 with the most significant difference exists in the right and left alagenion (outer curvature of the nose) and the right and left cheilion (corner of the mouth). The same non-symmetrical pattern is also observed with the left exocanthion more protrusive than the right.

As mentioned previously, these two skulls were the only skulls with reliable reference data. All other skulls presented faces that were digitally skewed or the data did not correlate along the same patterns as the previous two. Yet, we do not know why the skulls did this.

Discussion

Our data shows that, regardless of original sex of the skull, sexual dimorphism follows a similar pattern. From the two skulls, there is an average data set with 1.0146 for the right cheilion, 1.23435 for the left cheilion, 0.4373 for the labiale inferius, 0.48135 for the labiale superius, 0.7056 for the stomion, 2.3343 for the right alagenion, 1.89985 for the left alagenion, 0.67585 pronasale, 1.27465 for the subnasale, 0.45085 for the right endocanthion, 0.7205 for the left endocanthion, 0.5329 for the right exocanthion, 0.9708 for the left exocanthion, 0.5364 for the right ocular anterior, 0.5785 for the left ocular anterior. The graph (3.1) below shows a similar pattern in 3D differences of areas.

(3.1) Average 3D Difference

All the various distances match the same pattern as graphs 1.3 and 2.3 meaning sexual dimorphism ranges in skulls remains the same. This is of course in relation to collected averages in the TIVMI software reference. The usage of data from two skulls differing in biological sex shows that sexual dimorphism in facial reconstruction software remains similar in pattern with very close conclusions of measurement.

Conclusion

Sexual dimorphism in facial reconstruction software averages, as relating to the used TIVMI

software, exists. Using two skulls differing in biological sex, we found a similar pattern in averaged sex points with the greatest 3D difference in the right and left alagenion (outer curvature of the nose) and the right and left cheilion (corner of the mouth). The same non-symmetrical pattern is also observed with the left exocanthion more protrusive than the right. It is important to note that most TIVMI averages were collected from a French population and the skulls were taken from United States population samples.

While this is interesting, further research is needed to test these averages against a larger test

group. I suggest a future inquiry with a larger number of skulls with classes of both biologically determined sex and sexual expression between them. A further study into comparative sexual dimorphism between cultural races in a geopolitical group (example: comparative averages of sexual dimorphism between United States Caucasian, African, North Asian, South Asian, and Pacific Island groups) would also allow for a more well-known basis for sexual dimorphism. With a more well-known basis in human sexual dimorphism, we can understand from a biological perspective these averages in relation to human attractiveness or other sex-based inquiries.

Facial photos

Skull QINHN01_0112_Skull200 (female)

Anterior View (1.1)

Lateral View Left (1.2)

Lateral view Right (1.3) Skull QINHN01_0003_skull200 (male)

Anterior View

Lateral Left View

Lateral Right View

[1] Bruno, Dutailly. “http://projets.pacea.u-bordeaux.fr/TIVMI/index.php?page=kesako” 9/9/2014

[2] ‘’http://meshlab.sourceforge.net/’’ 9-9-2014

[3] Xanth D.G. Miller et al. “An Exploration of Sample Representativeness in Anthropometric Facial Comparison.” Journal of Forensic Science, July 2010. Volume 55, No 4. DOI 10.1111/j.1556-4029.2010.01425.x

[4] M. De Menezes et al. “Three dimensional analysis of labial morphology: effect of sex and age.” International Journal of Oral & maxillofacial surgery. 2011.

[5] Chiarella Sforza et al. “Age- and Sex-related changes in three-dimensional lip morphology.” Journal of Forensic Science. February, 2010.

[6] Sandra Anic-Milosevic, DDS et al. “Proportions in the upper lip-lower lip-chin area of the lower face as determined by photogrammetric method.” Journal of Cranio-Maxillo-Facial Surgery. 2010.

[7] Mohammed Bayome et al. “New three dimensional cephalometric analyses among adults with a skeletal class 1 pattern and normal occlusion.” The Korean Journal of Orthodontics. 2005.

[8] J. Wirthlin et al. “Comparison of facial morhpologies between adult Chinese and Houstonian Caucasian populations using three-dimensional imaging.” International Journal of Oral & Maxillofacial Surgery. 2013; 42:

[9] -1107. Page 5.

[10] J. Wirthlin et al. “Comparison of facial morhpologies between adult Chinese and Houstonian Caucasian populations using three-dimensional imaging.” International Journal of Oral & Maxillofacial Surgery. 2013; 42: 1100-1107. Page 7.