Semi-Supervised Dual-Threshold Contrastive Learning for Ultrasound Image Classification and Segmentation

2025-08-04

AI Summary

In semi-supervised contrastive learning for ultrasound image analysis, performance is severely hindered by erroneous pseudo-labeling—particularly overconfident early predictions—and the disjoint modeling of classification and segmentation tasks. To address these challenges, we propose Hermes, a dual-threshold contrastive learning framework. First, Hermes introduces a dual-threshold pseudo-label filtering mechanism to suppress error propagation. Second, it incorporates cross-task attention and saliency modules to enable shared feature representation between classification and segmentation. Third, it establishes cross-task consistency learning to align tumor-region representations and mitigate negative transfer. Evaluated on multiple public and private ultrasound datasets, Hermes achieves significant improvements over state-of-the-art methods under low-labeling regimes (≤20% annotated data), demonstrating superior robustness and generalization capability.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Segmentation
• Search and Optimization: Learning to Search

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
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Confidence-based pseudo-label selection usually generates overly confident yet incorrect predictions, due to the early misleadingness of model and overfitting inaccurate pseudo-labels in the learning process, which heavily degrades the performance of semi-supervised contrastive learning. Moreover, segmentation and classification tasks are treated independently and the affinity fails to be fully explored. To address these issues, we propose a novel semi-supervised dual-threshold contrastive learning strategy for ultrasound image classification and segmentation, named Hermes. This strategy combines the strengths of contrastive learning with semi-supervised learning, where the pseudo-labels assist contrastive learning by providing additional guidance. Specifically, an inter-task attention and saliency module is also developed to facilitate information sharing between the segmentation and classification tasks. Furthermore, an inter-task consistency learning strategy is designed to align tumor features across both tasks, avoiding negative transfer for reducing features discrepancy. To solve the lack of publicly available ultrasound datasets, we have collected the SZ-TUS dataset, a thyroid ultrasound image dataset. Extensive experiments on two public ultrasound datasets and one private dataset demonstrate that Hermes consistently outperforms several state-of-the-art methods across various semi-supervised settings.

Problem

• Overconfident incorrect pseudo-labels degrade semi-supervised learning
• Independent treatment of segmentation and classification limits performance
• Lack of public ultrasound datasets hinders model training

Innovation

• Dual-threshold contrastive learning for semi-supervised tasks
• Inter-task attention module for classification and segmentation
• Consistency learning to align tumor features