An era that artificial intelligence (AI) can diagnose dementia by analyzing the facial photo of patients is expected to arrive soon.
According to the Yomiuri Shimbun on Tuesday, a team of researchers from Tokyo University had more than 90% success rate of diagnosing Alzheimer's just by using AI system that analyzes the facial photos of patients. This is the first time that the diagnosis of dementia is made by using facial photos.
The study used data 121 men and women (an average of 81 years old) with mild cognitive impairment (MCI) or Alzheimer's and 117 men and women (an average of 76 years old) with normal cognitive function. The research team verified whether the AI system can distinguish between people with Alzheimer's and normal people using frontal, expressionless, facial photos. As a result, the AI system the accuracy rate of 93%. The accuracy rate was 91% when the AI system only analyzed the upper part of faces and 94% when analyzed the lower part of faces.
This study is a further development from the existing one that is based on the assumption, “People with reduced cognitive function tend to look older than they actually are.” The Tokyo University research team trained the AI on photos of people of Alzheimer's so that it can identify the micro-expressions on their faces.
“Early diagnosis of Alzheimer's is becoming more important in an aging society,” said Yumi Kameyama, a research associate at Tokyo University Hospital. “Facial analysis using AI will help early detection of Alzheimer's since it is safe and affordable.” The research team expects that AI analysis will help with the difficulty of detecting mild Alzheimer's.
Dementia refers to a condition, where it is difficult for an individual to cope with daily life due to impaired cognitive function caused by brain damage. It is not a disease itself but a general term. There are 80 to 90 causes of dementia, including Alzheimer’s disease and alcohol induced dementia. Alzheimer’s disease accounts for 50-80% of dementia cases.