Salcit Technologies, an India-based respiratory healthcare company, has recently partnered with the Google Research team to leverage Google’s Health Acoustic Representations (HeAR) technology in expanding the capabilities of their own bioacoustic AI technology, Swaasa. Swaasa utilizes HeAR to enhance research and improve the early detection of tuberculosis through analyzing cough sounds. This collaboration highlights the substantial potential of acoustic biomarkers in reshaping the diagnosis and monitoring of tuberculosis.
Sujay Kakarmath, a product manager at Google Research specializing in HeAR, emphasized the gravity of missed or late diagnoses of tuberculosis, describing them as tragedies and heartbreaks. He expressed gratitude for the role HeAR can play in revolutionizing the approach to tuberculosis detection and treatment.
Google’s HeAR technology was developed by training on a substantial set of 300 million pieces of audio data, derived from a diverse and anonymized dataset. The cough analysis model, specifically, was refined using about 100 million cough sounds. This extensive training allows HeAR to effectively identify patterns within health-related sounds, thus building a robust foundation for medical audio analysis applications.
Salcit Technologies has long utilized machine learning to advance early disease detection while addressing challenges of accessibility, affordability, and scalability. Their technology offers a location-independent, equipment-free option for respiratory health assessments. By integrating HeAR, Salcit aims to significantly widen the scope of tuberculosis screening throughout India.
This initiative has also garnered support from international health organizations, such as the United Nations’ Stop TB Partnership, which collaborates with tuberculosis experts and affected communities to eradicate TB by 2030. Zhi Zhen Qin, a digital health specialist with the Stop TB Partnership, commented on the groundbreaking potential of HeAR in transforming TB screening and detection, noting its accessibility and low-impact nature as critical advantages for reaching vulnerable populations.
Beyond this collaboration, there is a broader trend of utilizing sound and machine learning technologies in health diagnostics and monitoring. For instance, in 2022, EKO, a company specializing in smart stethoscopes, received FDA clearance for its Eko Murmur Analysis Software, which employs a machine learning algorithm to detect heart murmurs using heart sounds, phonocardiograms, and ECG signals in both adults and children.
Similarly, the Israeli health tech firm TytoCare secured significant growth funding for its AI-enabled TytoCare Home Smart Clinic. This innovation enables clinicians to perform remote exams using a suite of connected diagnostic devices, including an FDA-approved stethoscope that can analyze lung sounds to detect wheezing.
Another notable advancement comes from Canary Speech, which specializes in speech analysis software. Canary recently entered into a partnership with Microsoft to utilize AI technology in expanding its machine learning speech models for healthcare purposes. Canary’s technology detects irregularities in speech, aiming to address mental health issues, reduce healthcare costs, and scale up remote patient monitoring solutions.
The use of AI and sound analysis in healthcare continues to garner attention and development, emphasizing the potential of these technologies to revolutionize various aspects of medical diagnostics and patient care. Advances in these areas are likely to be highlighted and further discussed at the upcoming HIMSS AI in Healthcare Forum, scheduled for September 5-6 in Boston, where industry experts will gather to explore the integration of AI technologies in improving health outcomes and operational efficiencies within the healthcare sector.
#Salcit #partners #Google #technology #detect #disease #based #coughs
