Google's HeAR AI model could improve health diagnostics through bioacoustic analysis

It has been trained on 300 million audio samples

Image credit: Image Creator in Microsoft Bing/DALL.E and Canva

Google has introduced an innovative AI model called Health Acoustic Representations (HeAR) designed to detect diseases through audio analysis.

Here are a few highlights:

• This model has been trained on 300 million audio samples, including sounds like coughs, sneezes, and breaths. 

• By identifying patterns in these health-related sounds, HeAR can assist in early disease detection, making it a valuable tool for healthcare. 

• The model has demonstrated high performance even with less training data, which enhances its applicability in various healthcare settings. 

• An India-based company, Salcit Technologies, is leveraging HeAR to improve their product, Swaasa, which focuses on early detection of tuberculosis (TB) through cough analysis.