Indian Startup Salcit Technologies Partners with Google to Combat TB Using AI

Google’s HeAR bioacoustic foundation model is trained on approximately 300 million audio data samples, including roughly 100 million cough sounds.

In an effort to tackle the growing burden of tuberculosis (TB) in India, Hyderabad-based startup Salcit Technologies is collaborating with Google to utilize the tech giant's Health Acoustic Representations (HeAR) model for early TB detection through cough sound analysis.

This partnership was announced by Google on Tuesday in a blog post detailing the potential of the HeAR model.

Google’s HeAR bioacoustic foundation model, introduced in March 2024, is trained on approximately 300 million audio data samples, including roughly 100 million cough sounds.

The model is designed to assist researchers in developing tools that can analyze human sounds, such as coughs, to detect early signs of diseases such as TB.

Salcit Technologies aims to build on its existing product, Swaasa, using the HeAR model. Launched in 2020, Swaasa employs an AI-driven algorithm to assess lung abnormalities.

By integrating the HeAR model, Swaasa could significantly enhance its capabilities in detecting TB, making screening more accessible across India.

Expanding Access to Healthcare

Sujay Kakarmath, a product manager at Google Research, highlighted the accessibility of sound as a diagnostic tool compared to more traditional methods such as blood tests and imaging.

"HeAR can pick up chest x-ray findings, tuberculosis, and even detect Covid from cough sounds," Kakarmath said.

He emphasized that in areas with limited access to advanced medical resources, a healthcare professional equipped with a machine learning model and a smartphone could collect sound samples and provide crucial diagnostic information.

The Stop TB Partnership, an organization hosted by the United Nations and dedicated to eradicating TB by 2030, has expressed support for this approach.

Zhi Zhen Qin, a digital health specialist with the organization, stated that solutions such as HeAR could revolutionize TB screening and detection by offering a low-impact, accessible tool to those in need.

Google’s blog post also mentioned that the company is inviting researchers interested in exploring the potential of the HeAR model to request access to the HeAR API.

By continuing to develop diagnostic tools based on sound analysis, Google hopes to contribute to better health outcomes for communities worldwide, particularly in regions where access to traditional diagnostic methods is limited.