true77
Google’s AI arm is delving into the field of “bioacoustics,” a blend of biology and sound that offers insights into how the presence of pathogens affects human health. Our sounds can reveal valuable information about our well-being, and Google has developed an AI model to predict early signs of disease using sound signals.
How does Google’s bioacoustics AI work?
The AI model, called HeAR (Health Acoustic Representations), was trained on 300 million two-second audio samples of coughs, sniffles, sneezes, and breathing patterns. These audio clips were sourced from non-copyrighted content on platforms like YouTube, including recordings from a hospital in Zambia where patients were screened for tuberculosis. HeAR has been trained on 100 million cough sounds to detect tuberculosis.
Google has partnered with Salcit Technologies, an AI healthcare startup in India, to improve HeAR’s accuracy for tuberculosis and lung health screening. Salcit’s own AI model, Swaasa, allows users to submit a 10-second cough sample through a mobile app, with a 94 percent accuracy rate in identifying diseases.
Despite its potential, HeAR faces challenges such as background noise interference in audio samples. While the bioacoustics-based AI model is not yet market-ready, the concept of using AI and sound in medical diagnostics shows promise for the future.
Topics: Artificial Intelligence
