They enlisted a tech geek’s favourite gadget to help in the detection of a relatively common but sometimes tricky to diagnose neurological condition.
To most people, Raspberry Pi is a tasty dessert but tech wizards have used the fruit-named micro-computer, along with AI and a digital camera, to slash errors in facial palsy detection.
The real-time detection system identifies whether a patient has facial palsy with 98% accuracy. The deep learning tool can also detect the patient’s gender and age, according to the research published in BioMedInformatics.
Facial palsy is a neurological disorder that causes facial muscles on one side of the face to droop. It affects approximately 1 in 60 people worldwide during their lifetime.
Clinically it’s tricky to diagnose facial palsy because it’s symptoms can mimic other medical conditions such as stroke, tumor, HIV infection and multiple sclerosis.
Professor Javaan Chahl of the University of South Australia, said the diagnostic tool was 98% accurate in detecting facial palsy. Other research indicates that misdiagnosis happens in up to 20% of cases.
“Using computer vision systems to detect facial palsy could not only prevent misdiagnosis, but also save patients and medical specialists time, effort and cost,” he said.
Professor Chahl’s team at University of South Australia collaborated up with scientists at Middle Technical University in Baghdad, Iraq, to create the screening tool. The study authors suggested that it could be used as an auxiliary medical diagnostic tool for both clinical staff and patients.
“The patient’s use of this system at home in the diagnostic process reduces embarrassment, effort, time, and cost,” they said.
The tech behind the tool is both highly complex and rather accessible.
Raspberry Pi is a small, energy efficient computer loved by tech geeks and teachers for its versatility and affordability. A single board Raspberry Pi computer, around half the size of a standard smart phone, was installed with code to engage AI.
A Kisonli digital camera with a digital zoom was also attached to photograph the patient and display their image. Various deep learning codes were used including the Susan edge algorithm to identify the salient points of the edges and detect facial features.
Using a dataset of 26,000 images, containing 19,000 normal images and 1600 facial palsy images, researchers employed AI techniques to train computer vision systems to recognise the condition, differentiating them from healthy individuals.
The scientists then took photos of 20 patients with different degrees of facial palsy, using an algorithm to detect the condition in real time, as well as identifying their approximate age and gender.
People most at risk of developing facial palsy are usually aged between 30 and 45 years, pregnant women, diabetics, and those with a family history. The condition normally resolves spontaneously within six months.