Researchers from the Georgia Institute of Technology, led by research engineer Maia Gatlin, developed a device that can diagnose diarrheal diseases, like cholera and cancer, by listening to bathroom sounds. According to MailOnline, It uses artificial intelligence (AI) to look for subtle changes during defecation, urination, and flatulence.
The team built a database with hours of video and audio samples of human excretions from healthy and unhealthy individuals to prime the AI algorithm. They aim to find a non-invasive procedure to let people know whether they should get checked out or not. The invention comes as experts all over the world explore new ways of non-invasive diagnosis of cancer.
The S.H.A.R.T. Machine
The team named the new sensor the Synthetic Human Acoustic Reproduction Testing machine (S.H.A.R.T.). At the presentation during the annual Fluid Dynamics conference of the American Physical Society, the team said that the results of their study have not yet been published in a peer-reviewed journal but they are working on a device to pick up signals of deadly diarrheal diseases.
It is important to identify the source of an outbreak of a communicable disease as soon as possible. Researchers believe that their toilet sensor could help detect cholera in people of a given area by identifying its main symptom: diarrhea.
Gatlin and her team created the prototype device, which includes a microphone that records and listens as people use the toilet. New Atlas reported that using the AI algorithm running on an integrated microprocessor, the device can pinpoint the distinct audio signature of the loose, watery bowel movements associated with diarrhea.
They trained the AI using audio and video samples of excretion events obtained from online sources. Then they transformed it into a spectrogram that creates a visual representation of the sound.
Since each sound is associated with a certain type of event, the algorithm learns to distinguish which distinctive spectrogram features accompanied which sorts of excretion. When subsequently presented with spectrograms of other audio samples, the AI accurately identifies the corresponding event types. More importantly, it was able to single out diarrhea sounds even with background noises.
Combining Machine Learning and Sensors To Detect Diarrhea
According to Inverse, around 500,000 children die every year due to diarrheal diseases like cholera kill. It is the third leading cause of mortality in children worldwide. In fact, Haiti is experiencing an outbreak and resurgence of cholera, and ramping up the detection of such diseases will greatly help with treatment and prevention.
The team's goal of using the device is to combine machine learning models with inexpensive sensors and deploy them in the most susceptible areas. Gatlin said that classifying diarrheal events help them collect data that can help in those areas experiencing an outbreak.
They are looking forward to expanding their testing and although their work is still in its preliminary stage, they hope to make a very large impact using techniques that can deliver the hopeful capability for diagnosis.
Check out more news and information on Artificial Intelligence in Science Times.
