Medical Image Segmentation with Limited Supervision: A Review of Deep Network Models

TL;DR

This review discusses deep learning methods for medical image segmentation that operate effectively with limited supervision, addressing challenges like annotation scarcity and highlighting recent solutions and future research directions.

Contribution

It provides a comprehensive, up-to-date review of models and strategies designed for limited supervision in medical image segmentation, emphasizing recent advances and ongoing challenges.

Findings

Summarizes recent deep learning approaches for limited supervision in medical segmentation.

Highlights key challenges and potential future research directions.

Provides insights into methodologies that reduce annotation costs.

Abstract

Despite the remarkable performance of deep learning methods on various tasks, most cutting-edge models rely heavily on large-scale annotated training examples, which are often unavailable for clinical and health care tasks. The labeling costs for medical images are very high, especially in medical image segmentation, which typically requires intensive pixel/voxel-wise labeling. Therefore, the strong capability of learning and generalizing from limited supervision, including a limited amount of annotations, sparse annotations, and inaccurate annotations, is crucial for the successful application of deep learning models in medical image segmentation. However, due to its intrinsic difficulty, segmentation with limited supervision is challenging and specific model design and/or learning strategies are needed. In this paper, we provide a systematic and up-to-date review of the solutions above, with summaries and comments about the methodologies. We also highlight several problems in this field, discussed future directions observing further investigations.