A comprehensive survey on deep active learning in medical image analysis

TL;DR

The paper addresses the high annotation cost in medical image analysis by surveying deep active learning (AL) methods, especially their core components—informativeness evaluation and sampling strategies—and their integration with semi-supervised, self-supervised, and other label-efficient techniques. It introduces a comprehensive taxonomy of uncertainty- and representativeness-based AL approaches, discusses region-based and generative methods, and highlights practical considerations for medical tasks such as classification, segmentation, and reconstruction. The authors provide empirical evaluations across multiple datasets, demonstrating when uncertainty and representativeness strategies excel and how budget size shifts performance, while also sharing data and code to promote reproducibility. Finally, the survey outlines challenges and future directions, including uncertainty calibration, better representativeness metrics, weak annotations, advanced generative models, and foundation-model integration, underscoring the potential for impactful, cost-reducing AL methods in clinical practice.

Abstract

Deep learning has achieved widespread success in medical image analysis, leading to an increasing demand for large-scale expert-annotated medical image datasets. Yet, the high cost of annotating medical images severely hampers the development of deep learning in this field. To reduce annotation costs, active learning aims to select the most informative samples for annotation and train high-performance models with as few labeled samples as possible. In this survey, we review the core methods of active learning, including the evaluation of informativeness and sampling strategy. For the first time, we provide a detailed summary of the integration of active learning with other label-efficient techniques, such as semi-supervised, self-supervised learning, and so on. We also summarize active learning works that are specifically tailored to medical image analysis. Additionally, we conduct a thorough comparative analysis of the performance of different AL methods in medical image analysis with experiments. In the end, we offer our perspectives on the future trends and challenges of active learning and its applications in medical image analysis.

Figures (6)

• Figure 1: Illustration of the process of active learning.
• Figure 2: Overall framework of this survey.
• Figure 3: The taxonomy of uncertainty-based active learning.
• Figure 4: The taxonomy of representativeness-based active learning.
• Figure 5: The taxonomy of different sampling strategies in active learning.