Indian Startup Partners with Google to Transform TB Detection Using AI-Driven Cough Analysis

In a groundbreaking effort to tackle the rising burden of tuberculosis (TB) in India, Salcit Technologies, a Hyderabad-based startup, is set to enhance early TB detection through Google’s Health Acoustic Representations (HeAR) model. The HeAR model, introduced in March 2024, is trained on approximately 300 million audio data pieces, including 100 million cough sounds. This cutting-edge AI-driven technology allows for widespread TB screening across India by analyzing cough sounds to detect early signs of the disease.

Salcit Technologies plans to integrate the HeAR model into its product, Swaasa, which already uses AI algorithms to assess lung abnormalities. The initiative is supported by the Stop TB Partnership, a UN-hosted organization aiming to end TB by 2030, which views AI-powered acoustic analysis as a low-impact, accessible tool for TB screening.

Google is also inviting researchers to explore HeAR’s potential to further advance diagnostic tools globally.