New training method helps AI models handle messy, varied medical image data

Hospitals do not always have the opportunity to collect data in tidy, uniform batches. A clinic may have a handful of carefully labeled images from one scanner while holding thousands of unlabeled scans from other centers, each with different settings, patient mixes and imaging artifacts. That jumble makes a hard task—medical image segmentation—even harder still. Models trained under neat assumptions can stumble when deployed elsewhere, particularly on small, faint or low-contrast targets.

This post was originally published on this site