Case Studies

Centaur's mobile-first labeling tools enable us to quickly and conveniently gather results. We've accelerated our research by building and refining our ML models in near real time. Centaur.AI has been a crucial partner throughout.

Segmenting graft-versus-host skin disease is a complex and time consuming task. Centaur collected over 10 vetted opinions per image from their network, and combining those segmentations together produced amazingly high-quality annotations.
‍

We knew we couldn’t rely on individual neuroradiologists—or traditional, non-medical outsourced labeling services—to segment different kinds of hemorrhages in thousands of brain CT scans. Centaur tackled this problem quickly and accurately, enabling us to focus our time on building cutting-edge AI.

Any data scientist will love working with Centaur Labs. They provided us with rich information on labeler disagreement that is both critical to publishing results of our work and unlike any labeling vendor out there. Our project also had unique requirements for the population of labelers, and Centaur was able to recruit over 650 labelers that met those requirements.