DyCON: Dynamic Uncertainty-aware Consistency and Contrastive Learning for Semi-supervised Medical Image Segmentation

2025-04-06

AI Summary

In semi-supervised medical image segmentation, low 3D accuracy arises from high uncertainty in lesion regions and severe class imbalance. To address this, we propose a dynamic uncertainty-aware framework integrating consistency learning and contrastive learning. Our key contributions are: (1) Uncertainty-weighted Consistency Loss (UnCL), which dynamically amplifies supervision for challenging samples and high-uncertainty regions via adaptive uncertainty estimation; and (2) Focal Entropy-aware Contrastive Loss (FeCL), which jointly incorporates dual focal mechanisms and adaptive entropy modulation to balance global consistency and local discriminability. Evaluated on four major benchmarks—ISLES’22, BraTS’19, LA, and Pancreas—our method achieves significant improvements over state-of-the-art approaches, particularly enhancing robustness and accuracy for small lesions and ambiguous boundaries.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Segmentation
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Consent frameworks and practices on the web

Abstract

Semi-supervised learning in medical image segmentation leverages unlabeled data to reduce annotation burdens through consistency learning. However, current methods struggle with class imbalance and high uncertainty from pathology variations, leading to inaccurate segmentation in 3D medical images. To address these challenges, we present DyCON, a Dynamic Uncertainty-aware Consistency and Contrastive Learning framework that enhances the generalization of consistency methods with two complementary losses: Uncertainty-aware Consistency Loss (UnCL) and Focal Entropy-aware Contrastive Loss (FeCL). UnCL enforces global consistency by dynamically weighting the contribution of each voxel to the consistency loss based on its uncertainty, preserving high-uncertainty regions instead of filtering them out. Initially, UnCL prioritizes learning from uncertain voxels with lower penalties, encouraging the model to explore challenging regions. As training progress, the penalty shift towards confident voxels to refine predictions and ensure global consistency. Meanwhile, FeCL enhances local feature discrimination in imbalanced regions by introducing dual focal mechanisms and adaptive confidence adjustments into the contrastive principle. These mechanisms jointly prioritizes hard positives and negatives while focusing on uncertain sample pairs, effectively capturing subtle lesion variations under class imbalance. Extensive evaluations on four diverse medical image segmentation datasets (ISLES'22, BraTS'19, LA, Pancreas) show DyCON's superior performance against SOTA methods.

Problem

• Addresses class imbalance in semi-supervised medical image segmentation
• Reduces high uncertainty from pathology variations in 3D images
• Enhances consistency and contrastive learning for accurate segmentation

Innovation

• Dynamic uncertainty-aware consistency loss for global learning
• Focal entropy-aware contrastive loss for local discrimination
• Dual focal mechanisms for handling class imbalance