Dr Alison Ridel is a postdoctoral researcher at the Forensic Anthropology Research Centre (FARC, Department of Anatomy, University of Pretoria) as well as a 3D imaging and statistical trainer within Bakeng se Afrika. She tells us more about her PhD and her mobility to Belgium, where she learned new imaging and statistical skills.
Since the Bakeng se Afrika project started, I have had the great opportunity to share the knowledge acquired during my PhD and could organise training sessions with students. I was fortunate to learn and become proficient in three-dimensional (3D) imaging and in running advanced statistical analyses adapted to biological anthropology. I completed my PhD in 2018 with a thesis titled: An automated computer-assisted approximation of the nose in South Africans from CBCT (Cone Beam Computed Tomography) scans (available on ResearchGate), in which I focused on addressing problems in the reliability and validity of facial approximation methods in South Africa.
Facial approximations are frequently used by the South African Police Services (SAPS) as a means to help the identification of a missing person. However, the current methodology is subjective, unreliable, and invalid in the South African context, as it is mostly based on North American standards. In order to address these issues, the project aimed to develop a computer-based programme from which the nose, a particularly important feature of the face for identification, can be estimated on unknown South Africans. Based on the collection of 200 CBCT scans of modern, adult white and black South Africans, we could develop a reliable automated 3D prediction method of the nose. Furthermore, instead of using manual landmarking, this new method, implemented during my PhD, was based on automatic landmarking. Indeed, manual landmarking, used in several manual and semi-automated prediction guidelines for facial approximation, may render the analysis less repeatable due to observer subjectivity and, consequently, have an impact on the accuracy of the final facial approximation. In order to address this subjectivity and thereby improve facial approximations, we developed an automated 3D method based on an automatic dense landmarking procedure using non-rigid surface registration. The validation, precision and accuracy of the technique on the nose was then published in the Forensic Science International journal (if you cannot access the article, please contact me on Research Gate).
Furthermore, the methodology allowed us to show the presence of variation among and within black and white South African groups, and thus provide population- and sex-specific nose approximations, instead of resting on North American standards (work presented at the annual conference of the American Academy of Forensic Sciences).
I focussed only on the nose during my PhD, but the project is now expanding to include the rest of the face and other specific facial features through the ongoing and future work of Honours, MSc and PhD students. The goal is to develop a 3D automated prediction method creating a reliable and accurate facial approximation of an unknown person. During my PhD, a strong collaboration was developed with Prof Dirk Vandermeulen from the KU Leuven University in Belgium. His help, assistance and knowledge on the advanced statistical analyses and image processing methods was invaluable.
As the KU Leuven University is one of the three European institutions involved in the Bakeng se Afrika project, I was fortunate to spend 3 weeks at the Medical Imaging Research Center (MIRC) in September 2019.
During the first week of the mobility, I accompanied and conducted work with Sonè van der Walt, a PhD student under my supervision. Prof Dirk Vandermeulen, Dr Harry Matthews and I, could develop her PhD methodology by working on the robust extraction of anatomical faces from tomographic data (i.e., scans), but also investigating the method of placing landmarks, both on 2D and 3D surfaces of the same individual. The use of 2D or 3D anatomical landmarks allows for consistent analyses across studies as the landmarks are defined by specific anatomical structures and are therefore consistent across individuals. During our work, we also created a new module for the MeVislab software in order to measure absolute distances between planes.
The last two weeks of my mobility were more focused on my current postdoctoral project, which aim to evaluate the influence of ancestry and climatic adaptation on the morphology of the mid-facial skeleton, with a particular emphasis on developing standards for predicting mid-facial variation among modern black South Africans, for forensic implications. More precisely, by using micro-focus X-ray computed-tomography (micro-CT, a high-resolution scanning technique), this project will investigate mid-facial variation among Southern Africans but also other worldwide groups, such as Inuits, Northern and Southern Native Americans and Indigenous Australians.
During the mobility, with the help of Prof. Peter Claes and Dr. Harry Matthews, I could also learn how to use and develop in Meshmonk, an open source software defined as a toolbox producing a dense mesh of vertices across the entire surface, thus facilitating more comprehensive investigations of 3D shape variation. This software allows us to more accurately assess the biological shapes and better understand phenotype variation, genetic and developmental underpinnings, as well as the evolutionary history of modern humans.
Thanks to Bakeng se Afrika, this mobility at Leuven was an opportunity to acquire new competencies, that I could share with students back home which are beneficial for several of our research projects at the University of Pretoria. Of course, I also ate a lot of delicious waffles and Belgian chocolate!
Dr Alison Ridel
Postdoctoral researcher
Forensic Anthropology Research Centre
Department of Anatomy
University of Pretoria
Commentaires