Medical Audio Annotation for Speech AI

Audio clips can be leveraged to train predictive algorithms.

Our annotation platform is tailored specifically to audio clips, with features such as range selection, spectrogram integration, and motion data integration.

Use cases

Heart auscultation

Tag the presence of, intensity and time range of heart murmurs, rubs, clicks, S1, S2 and more.

Lung auscultation

Tag the presence of, intensity and time range of crackles, ronchi, wheezes, stridor, breath sounds and more.

Artery auscultation

Tag the presence of, intensity and time range of bruits and Korotkoff sounds.

Abdominal auscultation

Tag the presence of, intensity and time range of bowel sounds, rubs, or vascular bruits.

Clinical session recording

Tag intonation, emotion, symptoms and medications discussed in clinical interactions.

Annotation types

Classification

Range selection

Eko Health

Eko Health uses Centaur.ai to annotate lung sounds for its AI-enabled digital stethoscope

20,000 Lung sound recordings

120,000 Expert opinions

2 Week turnaround

.87 to .92 AUC

Eko Health and Centaur Labs

Trusted by AI leaders across healthcare

“The Centaur.ai platform provided labels at a scale that was 10x, or 20x, anything we had done by ourselves. Tremendous scale, tremendous throughput, and high quality labels.”

Daniel Barbosa
Machine Learning Engineer at Eko

"I was very impressed by the Centaur.ai approach to annotation. They provided us with excellent labels on what was a difficult task and noisy data. Centaur.ai understood what the issues were from the beginning, and were able to help steer the project for us."

Gareth Jones
Machine Learning Team Lead at Feebris