SoNUS AI

Sound-based Natural Understanding System

SoNUS was built to extract health data from audio collected by standard microphones

The type of health screening SoNUS AI provides depends on the placement and distance of the microphone in relation to the body. For cardiovascular analysis, for example, a microphone – such as a smartphone microphone – need simply be held against the chest.

Data collection

While our tech is now primarily deployed on smartphones, uniquely we built our own proprietary data collection devices from scratch. This was in order to guarantee entirely proprietary, custom and high quality data to train our AI models.

Research

While all our IP is owned by the company, and all research happens in-house, current team members have previously published widely in peer-reviewed journals on underlying principles and adjacent fields.

SoNUS AI healthscreening

Smallest system


Mental wellbeing analysis

Neurological analysis

Heart rate
variability algorithms

Heart rate extraction algorithms

Arrhythmia and murmur detection algorithms

Respiratory rate extraction algorithms

Expanded system


Cardiovascular analysis

Respiratory analysis

Gastrointestinal sounds detection algorithms

Cough detection algorithms

Medium-term state analysis

Gastrointestinal analysis

Limitless Expansion - Foundation Model

A single, previously collected audio file can be mined over and over again. This means that, as we expand the types of health analyses our engine is capable of, we can use the same original audio file to discover new health markers. Eventually we will be able to expand from our multi-model AI to the world’s first audio-to-health foundation model.