Researchers say the sound of a toilet flushing can detect the presence of a disease in a person. The study was conducted by researchers led by Maya Gatlin from the Georgia Institute of Technology in the United States. Gatlin described how a microphone sensor can detect intestinal diseases. Cholera is a bacterial disease that causes diarrhea that affects millions of people and causes about 150,000 deaths each year. Recognizing the potential for disease spread in the community provides health workers with early warning and improves the allocation of resources and assistance.
However, for obvious reasons, monitoring this and other intestinal disorders is a delicate matter. The fecal thesis: Using machine learning to detect diarrhea, Gatlin described how she and her team tested the technique, but using experimental sounds rather than reality. -World settings. Each audio sample of the output event has been turned into a spectrograph, essentially capturing the sound in the image. Different trips to the toilet produce different characteristics in tone and in the spectrogram. For example, urination produces a constant sound, while stool may have a single sound.
By contrast, diarrhea is more random. The spectrogram images were sent to a machine learning algorithm that learned to classify each event based on its properties. The algorithm’s performance is tested on data with and without background noise to ensure it learns the correct noise characteristics regardless of the sensor’s environment. In the future, Gatlin and his colleagues plan to collect real-world audio data so their machine learning model can be adapted to a variety of bathroom environments. “We hope this small sensor can be used in areas where outbreaks of cholera pose an ongoing risk,” Gatlin said.
“The sensor could also be used in disaster areas (where water pollution leads to the spread of pathogens in the water) or even in aged care facilities to automatically monitor patients’ stool,” he added.
Site Title 