Medical Image Segmentation

Our unique approach aggregates multiple opinions to generate a single precise annotation (across different regions of interest). We weigh all annotations based on individual skill level and mark all annotated regions with a confidence score.

Polygon Annotation

Precise segmentation of findings in medical imagery (like brain hemorrhages, aneurysms, and lung nodules)

3D Annotations

Segmentation across all frames of 3D imagery, including DICOM and NIfTI formats, and modalities like CT and MRI

Box Annotation

Bounding boxes around one or multiple findings


Measurement of anatomical features

Medical Image Segmentation


"Trace the pneumothorax."

  • Polygon segmentation of pneumothorax (collapsed lung) in chest x-rays

"Segment the hemorrhage area."

  • Precise segmentations of areas of bleed within 3D brain CT scans
  • DICOM imagery included multiple contrast levels for maximum visibility

"Draw a box around the polyp(s)."

  • Box annotation of all polyps present in colonoscopy footage
Want to see more of our work? View Projects →

About Centaur Labs

Centaur Labs focuses exclusively on medical data annotation to enable breakthroughs in medical AI. Our unique approach combines multiple opinions from our network of medical experts to deliver accurate segmentations.

Get started with Centaur Labs today