Instance-Aware Pseudo-Labeling and Class-Focused Contrastive Learning for Weakly Supervised Domain Adaptive Segmentation of Electron Microscopy

2025-10-18

AI Summary

Mitochondrial instance segmentation in electron microscopy (EM) images suffers from high annotation costs and heavy reliance on expert knowledge. To address this, we propose a weakly supervised domain adaptation (WDA) method leveraging sparse point annotations. Our approach features: (1) an instance-aware pseudo-label selection strategy that exploits a center detection task to generate semantically reliable and diverse pseudo-labels; and (2) a class-focused cross-domain contrastive learning mechanism that enhances intra-class consistency of target-domain instances while suppressing inter-class confusion in feature space. The framework jointly optimizes segmentation and center detection via multi-task learning, and improves unlabeled data utilization through integrated cross-teaching and self-training. Extensive experiments on multiple EM datasets demonstrate that our method significantly outperforms existing unsupervised and weakly supervised domain adaptation approaches, substantially narrowing the performance gap with fully supervised baselines—while also achieving notable gains under pure unsupervised domain adaptation (UDA) settings.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Segmentation
• Application Domains: Software Engineering

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Annotation-efficient segmentation of the numerous mitochondria instances from various electron microscopy (EM) images is highly valuable for biological and neuroscience research. Although unsupervised domain adaptation (UDA) methods can help mitigate domain shifts and reduce the high costs of annotating each domain, they typically have relatively low performance in practical applications. Thus, we investigate weakly supervised domain adaptation (WDA) that utilizes additional sparse point labels on the target domain, which require minimal annotation effort and minimal expert knowledge. To take full use of the incomplete and imprecise point annotations, we introduce a multitask learning framework that jointly conducts segmentation and center detection with a novel cross-teaching mechanism and class-focused cross-domain contrastive learning. While leveraging unlabeled image regions is essential, we introduce segmentation self-training with a novel instance-aware pseudo-label (IPL) selection strategy. Unlike existing methods that typically rely on pixel-wise pseudo-label filtering, the IPL semantically selects reliable and diverse pseudo-labels with the help of the detection task. Comprehensive validations and comparisons on challenging datasets demonstrate that our method outperforms existing UDA and WDA methods, significantly narrowing the performance gap with the supervised upper bound. Furthermore, under the UDA setting, our method also achieves substantial improvements over other UDA techniques.

Problem

• Mitochondria instance segmentation from EM images
• Weakly supervised domain adaptation with sparse labels
• Improving pseudo-label reliability through instance-aware selection

Innovation

• Instance-aware pseudo-labeling for reliable segmentation
• Class-focused contrastive learning for domain adaptation
• Cross-teaching mechanism integrating segmentation and detection