LG-NuSegHop: A Local-to-Global Self-Supervised Pipeline For Nuclei Instance Segmentation

2025-11-07

AI Summary

Nuclear instance segmentation in histopathological images relies heavily on expert annotations, resulting in high labeling costs and poor generalization across organs and imaging domains. To address this, we propose the first fully unsupervised self-supervised framework that integrates biological priors with data-driven design. Our method comprises three key components: (1) generating interpretable pseudo-labels via localized image perturbations; (2) introducing NuSegHop, a robust feature extractor leveraging hierarchical patch-based representation learning; and (3) applying a global optimization strategy for end-to-end nuclear segmentation. Crucially, the entire pipeline requires no domain adaptation, external supervision, or manual intervention, and each module is clinically interpretable. Evaluated on three public benchmark datasets, our approach surpasses existing unsupervised and weakly supervised methods, matching the performance of state-of-the-art fully supervised models while demonstrating significantly improved cross-domain generalizability and clinical deployability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Nuclei segmentation is the cornerstone task in histology image reading, shedding light on the underlying molecular patterns and leading to disease or cancer diagnosis. Yet, it is a laborious task that requires expertise from trained physicians. The large nuclei variability across different organ tissues and acquisition processes challenges the automation of this task. On the other hand, data annotations are expensive to obtain, and thus, Deep Learning (DL) models are challenged to generalize to unseen organs or different domains. This work proposes Local-to-Global NuSegHop (LG-NuSegHop), a self-supervised pipeline developed on prior knowledge of the problem and molecular biology. There are three distinct modules: (1) a set of local processing operations to generate a pseudolabel, (2) NuSegHop a novel data-driven feature extraction model and (3) a set of global operations to post-process the predictions of NuSegHop. Notably, even though the proposed pipeline uses { no manually annotated training data} or domain adaptation, it maintains a good generalization performance on other datasets. Experiments in three publicly available datasets show that our method outperforms other self-supervised and weakly supervised methods while having a competitive standing among fully supervised methods. Remarkably, every module within LG-NuSegHop is transparent and explainable to physicians.

Problem

• Automating laborious nuclei segmentation without manual annotations
• Addressing large nuclei variability across tissues and domains
• Developing transparent self-supervised pipeline for medical generalization

Innovation

• Self-supervised pipeline without manual annotations
• Local-to-global processing for nuclei segmentation
• Transparent and explainable modules for physicians