Google develops an intelligent model for early detection of diseases via sound
Google has revealed the development of a new artificial intelligence model called HeAR, which aims to help in the early detection of some diseases using audio data.
The new model relies on analyzing human sounds, such as coughing, to predict early signs of certain diseases, including tuberculosis. The importance of this technology lies in the speed of diagnosis, which is a crucial factor in treating serious diseases.
The HeAR model was trained using 300 million audio clips, 100 million of which were cough sounds. This model outperforms its peers despite using less training data.
Interestingly, this technology can be integrated into mobile phones, opening up vast possibilities for its use in remote areas far from health facilities. Instead of relying on expensive medical imaging devices, the smartphone microphone can be used to conduct initial examinations.
Related > New blood test could diagnose Alzheimer's disease in its early stages
Google has partnered with Indian respiratory healthcare company Salcit Technologies, which has a proprietary model called Swaasa that uses cough sounds to assess a patient's lung condition.
Despite the anticipated challenges in convincing doctors to use this technology, the support of prestigious organizations such as the UN Stop TB Partnership enhances its credibility.
