Learning to Segment Medical Images with Scribble-Supervision Alone
This addresses the challenge of reducing annotation burden for medical professionals in segmentation tasks, though it is incremental as it builds on existing weak supervision methods.
The paper tackles the problem of training medical image segmentation networks using only scribble annotations instead of full pixel-wise labels, which are time-consuming for medical professionals to produce. The result shows that networks trained on scribbles achieve Dice scores with only a small degradation of 2.9% for cardiac and 4.5% for prostate datasets compared to full annotations.
Semantic segmentation of medical images is a crucial step for the quantification of healthy anatomy and diseases alike. The majority of the current state-of-the-art segmentation algorithms are based on deep neural networks and rely on large datasets with full pixel-wise annotations. Producing such annotations can often only be done by medical professionals and requires large amounts of valuable time. Training a medical image segmentation network with weak annotations remains a relatively unexplored topic. In this work we investigate training strategies to learn the parameters of a pixel-wise segmentation network from scribble annotations alone. We evaluate the techniques on public cardiac (ACDC) and prostate (NCI-ISBI) segmentation datasets. We find that the networks trained on scribbles suffer from a remarkably small degradation in Dice of only 2.9% (cardiac) and 4.5% (prostate) with respect to a network trained on full annotations.