New AI-based Facial Screening Tool for Detecting Stroke

TapTechNews June 20th news, a team of biomedical engineers from Royal Melbourne Institute of Technology (RMIT) and Guilherme Camargo de Oliveira, a Ph.D. student from the State University of São Paulo, have jointly developed an AI-based facial screening tool. Medical staff only need to use a smartphone to determine whether a patient has had a stroke within a few seconds.

The tool has an accuracy rate of 82% in detecting stroke, but it will not replace comprehensive clinical diagnostic tests for stroke, and is only used to help identify those in need of treatment more quickly. Professor Kumar from the RMIT School of Biomedical Engineering said that early detection and timely treatment of stroke can significantly improve recovery results and reduce the risk of long-term disability.

The tool deeply integrates AI with facial recognition technology and detects stroke by analyzing facial symmetry and specific muscle movements (called action units). The Facial Action Coding System (FACS) used by the tool was originally developed in the 1970s and classifies facial movements through the contraction or relaxation of facial muscles, providing a detailed framework for analyzing facial expressions.

A team member, de Oliveira, said that one of the key parameters affecting stroke patients is that their facial muscles usually become unilateral movement, so the muscle performance on both sides of the face is different. The core of this tool is to detect whether there will be a change in asymmetry when the face smiles. The study used video clips of facial expression checks of 14 post-stroke patients and 11 healthy controls.

The team plans to work with healthcare providers to develop this tool into a mobile application in order to be able to detect other neurological diseases that affect facial expressions.

TapTechNews attached the original website of the paper:

https://www.sciencedirect.com/science/article/abs/pii/S0169260724001913