Facial Recognition Helps Predict When ICU Patients Are At Risk

By Fred Donovan

June 03, 2019 - Japanese researchers have successfully incorporated facial recognition technology into an automated system that can predict when intensive care unit (ICU) patients are at high risk of unsafe behavior, such as removing their breathing tubes.

The automated risk detection tool can continuously monitor a patient’s safety and ease limitations associated with constrained hospital resources.

"Using images we had taken of a patient's face and eyes, we were able to train computer systems to recognize high-risk arm movement," said Akane Sato from Yokohama City University Hospital, Japan, who led the research.

Dig Deeper

Ochsner LSU Health Monroe Launches Telehealth ICU Program
Machine Learning Could Reduce Lab Testing for ICU Patients

"We were surprised about the high degree of accuracy that we achieved, which shows that this new technology has the potential to be a useful tool for improving patient safety and is the first step for a smart ICU which is planned in our hospital," Sato said.

ICU patients are often sedated to lessen pain and anxiety, allow for invasive procedures, and improve patient safety. However, providing the right level of sedation can be challenging. Patients who do not receive the correct level of sedation are more at risk for accidentally removing invasive devices, Sato related.

The study, conducted at the Yokohama City University Hospital between June and October 2018, involved 24 postoperative patients admitted to the hospital’s ICU.

The system uses photos taken by a camera mounted on the ceiling above the patients’ beds. The researchers analyzed 300 hours of data to find daytime images of patients facing the camera in a position that showed their face and eyes clearly.

The researchers then used artificial intelligence to analyze 99 images. The proof-of-concept system was able to alert for high-risk behaviors with a 75 percent accuracy rate.

"Various situations can put patients at risk, so our next step is to include additional high-risk situations in our analysis and to develop an alert function to warn healthcare professionals of risky behavior. Our end goal is to combine various sensing data such as vital signs with our images to develop a fully automated risk prediction system," said Sato.

The researchers identified several limitations with the system, including the need for more images of patients in different positions to improve the generalizability in real life. They also noted that monitoring the patient's consciousness may improve the accuracy in distinguishing between high-risk behavior and voluntary movement.

Latest News

Facial Recognition Helps Predict When ICU Patients Are At Risk

Japanese researchers have successfully incorporated facial recognition technology into an automated system that can predict when ICU patients are at high risk of unsafe behavior.

Next in Latest News