Low-Contrast-Enhanced Contrastive Learning for Semi-Supervised Endoscopic Image Segmentation

2024-12-03

AI Summary

Endoscopic image segmentation faces challenges due to low tumor–normal tissue contrast, ambiguous boundaries, and scarce pixel-level annotations. To address these issues, we propose LoCo, a semi-supervised framework. Its key contributions are: (1) Low-Contrast Contrastive Learning (LCC), the first contrastive learning paradigm tailored for low-contrast medical images, integrating Inter-Class Enhancement (ICE) and Boundary Contrastive Enhancement (BCE) to jointly improve feature discriminability and boundary localization; and (2) Confidence-Dynamic Filtering (CDF), a pseudo-label selection mechanism that adaptively thresholds confidence scores per class—particularly beneficial for under-represented foreground categories—to enhance pseudo-label reliability. Evaluated on two public endoscopy benchmarks and a large private clinical dataset, LoCo achieves state-of-the-art performance across all settings. The implementation is publicly available.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Semi-Supervised Learning
• 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
• Security and Privacy: Browser and end-point security

Abstract

The segmentation of endoscopic images plays a vital role in computer-aided diagnosis and treatment. The advancements in deep learning have led to the employment of numerous models for endoscopic tumor segmentation, achieving promising segmentation performance. Despite recent advancements, precise segmentation remains challenging due to limited annotations and the issue of low contrast. To address these issues, we propose a novel semi-supervised segmentation framework termed LoCo via low-contrast-enhanced contrastive learning (LCC). This innovative approach effectively harnesses the vast amounts of unlabeled data available for endoscopic image segmentation, improving both accuracy and robustness in the segmentation process. Specifically, LCC incorporates two advanced strategies to enhance the distinctiveness of low-contrast pixels: inter-class contrast enhancement (ICE) and boundary contrast enhancement (BCE), enabling models to segment low-contrast pixels among malignant tumors, benign tumors, and normal tissues. Additionally, a confidence-based dynamic filter (CDF) is designed for pseudo-label selection, enhancing the utilization of generated pseudo-labels for unlabeled data with a specific focus on minority classes. Extensive experiments conducted on two public datasets, as well as a large proprietary dataset collected over three years, demonstrate that LoCo achieves state-of-the-art results, significantly outperforming previous methods. The source code of LoCo is available at the URL of href{https://github.com/AnoK3111/LoCo}{https://github.com/AnoK3111/LoCo}.

Problem

• Endoscopic Image Analysis
• Tumor Recognition
• Limited Annotated Data

Innovation

• Local Contrast Learning
• Unsupervised Image Utilization
• Credibility-based Filtering Mechanism