eCPD - Cliniface - Unlocking Facial Clues
Date & Time
Monday, 11 May 2020
- 5:00 pm
Speaker: Dr Richard Palmer, Curtin University
Richard is a research fellow in the School of Earth and Planetary Sciences at Curtin University. He gained first class honours in Computer Science before receiving his doctorate in 2016 which concerned investigations into new techniques of fusing multi-modal (2D and 3D) imagery for pattern recognition and machine learning. Since then, he has been working closely with colleagues from the Department of Health and around the world to design and develop Cliniface – a software application to support clinicians and researchers in their investigations of the human facial phenotype for rare disease diagnosis.
Presentation: The Human Phenotype Ontology (HPO) is a structured set of well-defined terms to describe phenotypic abnormalities (i.e. disease indicators). HPO terms are widely used in the precision medicine and rare diseases communities for developing computational tools for diagnostic support and to monitor treatment outcomes. Cliniface unlocks the power of facial clues.
Recognising facial clues (phenotypes) requires clinical expertise about distinct vs subtle facial characteristics, and is challenged by overlapping facial phenotypes. The main challenges are:
- Using direct anthropometry measurements can be difficult and inaccurate to perform on young children;
- If measurements have to be revisited or are not taken, the patient has to return to hospital;
- There are a number of facial phenotypes which are highly subjective.
- Facial phenotypes may vary depending on ethnicity.
Cliniface is a desktop software application that gives users the ability to visualise and analyse 3D facial images captured by medical 3D imaging hardware. Cliniface delivers annotation of facial landmarks onto 3D surfaces using well known spatial methods. Those facial landmarks can be detected semi-automatically and can be adjusted without the requirement for the patient to sit still during a consultation, annotation and measurements can occur during or after a consultation, on pre-existing images, or on images taken in remote locations to reduce the need to travel.
Su Ling Meimaris