Learning to Segment Medical Images from Few-Shot Sparse Labels

TL;DR

This paper introduces a meta-learning approach for medical image segmentation that effectively learns from sparse labels in few-shot scenarios, reducing annotation costs while maintaining high accuracy.

Contribution

It presents a novel MAML-based method that trains on sparse labels and generalizes to dense label predictions in medical imaging, especially when target domains differ significantly.

Findings

Achieves Jaccard scores comparable to dense labels with less than 2% labeled pixels.

Effective in domain shifts with high differences between source and target datasets.

Outperforms existing methods in few-shot medical image segmentation scenarios.

Abstract

In this paper, we propose a novel approach for few-shot semantic segmentation with sparse labeled images. We investigate the effectiveness of our method, which is based on the Model-Agnostic Meta-Learning (MAML) algorithm, in the medical scenario, where the use of sparse labeling and few-shot can alleviate the cost of producing new annotated datasets. Our method uses sparse labels in the meta-training and dense labels in the meta-test, thus making the model learn to predict dense labels from sparse ones. We conducted experiments with four Chest X-Ray datasets to evaluate two types of annotations (grid and points). The results show that our method is the most suitable when the target domain highly differs from source domains, achieving Jaccard scores comparable to dense labels, using less than 2% of the pixels of an image with labels in few-shot scenarios.