Smart Toilet that Checks Health Statu
The Stanford University Graduate School of Medicine has developed a system that checks the health of urine and stool by adding gadgets and sensors to the toilet. Announced. With the camera images and machine learning provided, it is possible to identify the user based on the pattern of the anus and check the health status from urine components, flow rate, and stool shape with the same accuracy as a doctor.
The research team installed pressure sensors, motion sensors, urinalysis test strips, and various cameras on the toilet seat of the Western-style toilet.
Smart toilet system developed by the research team. Users can be authenticated with fingerprints and anus, and health information obtained from urine and stool can be sent to a cloud service (images quoted from papers)
First, the pressure sensor detects that the user is sitting on the toilet seat. When urination is detected by the motion sensor, the position of the test strip for urine component analysis is adjusted. Urine flow is photographed using two small high-speed cameras (GoPro HERO7), and the flow rate is estimated by image processing.
For stool, two doctors tagged the stool image data set (2362 sheets) based on indices classified into seven stages based on shape and hardness, and trained by deep learning. When taking photos of the stool with a camera attached to the toilet seat and classifying it using a learning model, the doctor was able to demonstrate the same diagnostic accuracy as a doctor.
A fingerprint authentication sensor is mounted on the toilet lever to link the health information obtained from urine and stool with the user. However, the user does not always flush the toilet, and the flushing user may be another person. Therefore, in addition to fingerprints, “anal wrinkles” were adopted as biometric authentication. Users can be identified by comparing the taken anus with machine learning (template matching).
User identification by template matching of anal image
According to the research team, the anal image is only used for recognition, and no one will see the image.
The research team hopes that checking excrement for health will lead to early detection of digestive disorders.
Since the current system can only be installed on a Western-style toilet, the team has also developed an application that can determine the status of the flight simply by shooting the flight with a smartphone.
In the future, it will incorporate other useful testing methods and conduct large-scale clinical research.