Research project
3DFaceAtHome
Using AI to make accurate facial 3D models from selfies images acquired with a mobile app at home
Active
April 2024
April 2026
3DfaceAtHome aims to develop and test a system to obtain accurate facial 3D models from end users' mobile devices in real world conditions, such as one's smartphone at home. The goal is to reach the precision required to measure facial features relevant to medical conditions affecting the face. In the future, this project will facilitate genetic studies of facial differences and enable clinical translation of such a system in related medical fields.
The system consists of a purpose-made mobile application guiding the user to remotely capture selfies images by themselves. A fully personalised AI system is then used to obtain the 3D shape of the user’s face. To evaluate the system, a dataset including examples of facial differences will be acquired from volunteer participants. It will contain a set of facial images and a ground-truth 3D facial scan collected in a lab setting, and optionally a set of facial images acquired by the volunteers themselves at home using the mobile application.
Get involved!
To help with this research, we will soon need the help of volunteers to contribute facial images. The data acquisition will be done at the University of Dundee and University of Glasgow between September and November 2025.
Criteria to participate in this project are:
- Being at least 18 years old and in capacity to consent
- Willingness to show your face from the neck up to the hair (included), and from ear to ear on the sides
- Absence of facial hair masking the skin such as beard or mustache
- Access to a device that can run our mobile application (Android and iOS based smartphones should work)
- Able to attend one of our acquisition session at the University of Dundee or University of Glasgow
The data collected include a 3D facial scan and two sets of facial images which will only be used or shared according to your choices. See the participant information sheet, privacy notice and consent form for more information on how we will acquire, process and share your data:
Dataset Access
When the dataset will be available, researchers will be able to access the data of volunteers that have agreed to share it. Details will be made available once the dataset is finalised.
Funders
The project is funded by EPSRC (ref. EP/X036642/1) as a New Investigator Award for Dr Ludovic Magerand.
This is a collaboration with Prof Peter Mossey (University of Dundee, School of Dentistry) and Prof Ashraf Ayoub (University of Glasgow, School of Dentistry).