How applying new research developments changes the face of unidentified human remains.
“I’ve been looking at your face more closely in the last half an hour than you probably have in years,” Dr Susan Hayes says.
From the positioning of the eyes to the skin tone and soft-tissue features, Dr Hayes looks at every face with the eye of an academic anatomist, physical anthropologist and accomplished artist.
Her intimate understanding of facial features, based on years of scientific research in the field of facial anthropology, now makes it possible to put a face to remains excavated at archaeological sites or to help police identify missing persons.
Dr Hayes is an Honorary Senior Research Fellow with the University of Wollongong’s Centre for Archaeological Science and her expertise is facial anthropology, which blends newly published scientific data about the skull and its soft tissues with the technical skills of artistic depiction to create evidence-based facial approximations from skulls.
Because I am always working with the skull of a deceased individual, then they are always the remains of someone who once lived, dreamed, had likes and dislikes - in essence lived all that it is to be human.Dr Susan Hayes
While her work is putting excavations on the map and allowing new, evidence-based insights into human history, it’s also challenging conventional wisdom and long-used methods of imagining past peoples.
“Because I am always working with the skull of a deceased individual, then they are always the remains of someone who once lived, dreamed, had likes and dislikes – in essence lived all that it is to be human,” Dr Hayes says.
“For the first time in over a century it is now possible to estimate a person’s face with scientific justification and a high degree of methodological transparency. Yet others – both artists and scientists, professionals and amateurs – overwhelmingly rely on one or two of the popular forensic science handbooks, and that’s pretty much the extent of their research engagement.
“Unfortunately, over the past 10 years or so, nearly all of the methods illustrated in these books have been shown to be, well, wrong. The deceased deserve a bit better than this.”
A nod to ancient history
In 2002, the remains of a woman were found in the Tham Lod rock shelter in north-west Thailand. She was between 25 and 35 years old. And she had been dead for more than 13,600 years.
She is likely a descendent from the first humans to ever colonise South-East Asia. Using a range of the newly developed facial approximation techniques, a team of researchers that included an archaeologist, anatomist, physical anthropologist and a dental specialist have now given her a face.
To assess this bias, Dr Hayes compared their estimated face of the Tham Lod woman with a series of facial measurements – such as the space between the eyes, the length of the nose – with the average facial dimensions of 720 contemporary women living in 25 different countries across three continents.
They wanted to figure out whether it is really possible, based on the new methods, to recreate the unique face of a woman whose face is neither recent nor European in appearance – and it turns out that it is.
Everybody wants to know where we come from. This should be done as scientifically as possible, with the least possible interpretation.Dr Gerrit van den Bergh
“We were worried that these skull-soft tissue relationships, because they are the averages from mostly European populations, might overwrite the face of a woman from the Late Pleistocene of Thailand.
“Which is a place and time that is not recent, not European. We were basically worried that our methods would quite literally colonise the face. But when we compared our results, statistically, with facial dimensions derived from women all over the world, we found that the face we had estimated was still quite distinct, still conformed to what we know about people living in the Late Pleistocene.
“What was most gratifying is that she has more in common with today’s East Asian woman, and very little in common with modern European women.”
No face is the same
What’s important from Dr Hayes’s perspective is that the image is a depiction based on the statistical averages of a range of skull-soft tissue relationships – and so the results are always going to be a kind of a composition of average features.
But, as she likes to say, “no one is average and neither are their skulls”. Rather than a portrait of what this particular woman looked like, we get a snapshot of the ancient features of a young woman who lived in this area.
“There’s an awful lot about the face that we don’t know, which is quite weird given that we are so focused on our faces,” Dr Hayes says.
Dr Gerrit van den Bergh, from the Centre for Archaeological Science at UOW, has worked on excavating the remains of Homo floresiensis, the tiny species of human, affectionately dubbed ‘Hobbit’, which stood at just one metre tall and lived on the Indonesian island of Flores.
Facial approximations of a woman from Tham Lod, Thailan. Images: Hayes, Shoocongdej et al 2017/Antiquity.
He says facial approximation is a valuable tool for understanding and exhibiting human history.
“Everybody wants to know where we come from,” Dr van den Bergh says. “This should be done as scientifically as possible, with the least possible interpretation. For example, Neanderthals were initially depicted as violent brutes, an image that is still prevalent. Over the last decade it became clear that that image was totally wrong.”
Through the application of robust and validated skull-soft tissue relationships, archaeologists can see that the methods are applied consistently, and the resulting facial approximation can be verified.
“So, it is the only way we have to reliably estimate how ancient humans might have looked,” Dr van den Bergh says.
An approximate guide to faces of the past
Reliably estimating a face from bones involves two main approaches: two-dimensional and three-dimensional. The 2D approach uses a photograph or outline of the skull and facial features are sketched over the top.
The 3D approach is the oldest and still most popular method, where clay is sculpted over a cast of the skull. A more modern variation of the 3D method is to use computer rendering software to create the image over the skull.
But whether it’s drawing, sculpting or, as Dr Hayes does, using computer graphics to create a virtual face, it’s the methods rather than the medium that has the greatest impact on the validity of the results.
It’s a tough gig to create soft features from bony structures; turning the lifeless into the life-like. The correlations between bony features of a skull and soft tissue on the face are complicated and can tell you nothing about how features such as ears and hair should look.
For the first time in over a century it is now possible to estimate a person’s face with scientific justification and a high degree of methodological transparency.Dr Susan Hayes
This is where data on average features in human populations comes in, and also explains why facial anthropologists prefer the terms “approximation” or “estimation” over “facial reconstruction”.
“The shape of the bones dictates the shape of the face quite strongly. But there are a whole lot of other relationships you apply at the same time,” Dr Hayes says. “I don’t know what someone is going to look like when I see a skull. It takes me a long time to go through the whole process and get there.”
Despite the highly technical nature of this work, it can be very personal. Working with skulls is a constant reminder of the person who once lived, and the approximation method slowly uncovers the humanity behind a collection of bones.
“These are the remains of a person, so it is very much like someone new has entered the lab and it will take time to get to know them,” Dr Hayes says. At the same time, Hayes keeps a respectful distance, recognising that this person already had an identity, and this true identity is not something we can know or create for the individual.
“This is, in part, why I never refer to the person I’m working with by name. I keep to their case or catalogue number. I know that they already have a name. I don’t happen to know it, and it is very likely I will never know it … but that doesn’t mean I can overwrite their history by naming them.”
The Belanglo State Forest murders: giving victims of crime an identity
Facial approximation not only helps uncover the faces of ancient humans, but those of the more recently deceased as well. In forensic facial approximation, a face is built on the skull of an unidentified victim of a crime. Broadcasting the image is hoped to trigger recognition in those who knew the person.
While the number of unidentified bodies in Australia is small, there is both a legal requirement and social right for victims to be identified. This is where more difficult methods such as facial approximation can come into play.
In August 2010, the skeletal remains of a young woman aged between 15 and 25 were found in Australia’s Belanglo State Forest. A year later, the New South Wales Police requested a facial approximation. It turns out that the estimation fitted well with two photos of the victim, although there were noticeable differences.
“It would have probably popped up on automated face recognition, but might have disrupted family recognition because how we recognise people that we know is not metric, it’s certain characteristics,” Dr Hayes says.
“Had she been on the missing person’s database, we would, I think, have recognised her. But she wasn’t.”
Respecting the past and those passed
Dr Hayes believes the main function of facial approximation in forensics is to create a “media bump”. The interest surrounding the science itself is enough to get people interested, and create new leads, and this may be more effective than actual recognition of the face itself.
Working on a forensic case brings a new kind of connection to the person. While Hayes was working on the facial approximation, there were people who still did not know this woman had died.
“She was still a living person in the active memories of living people, people who were going about their lives at exactly the same time as I was going about measuring her skull.”
Handling the skull of a recently deceased person can be far more difficult emotionally than that of a person who died thousands of years ago, and Dr Hayes doesn’t like to know the background to a forensic case unless it impacts on facial appearance.
But she says it also feels worthwhile to help with identification, not just for the families and friends, but also the police officers and forensic specialists who have often worked on the case for a long time.
“We all want to see the person identified, and we all want closure.”