MUSE: Multi-Scale Dense Self-Distillation for Nucleus Detection and Classification

2025-11-07

AI Summary

Nuclear detection and classification (NDC) heavily relies on dense manual annotations and struggles to leverage abundant unlabeled histopathological images. To address this, we propose MUSE, a self-supervised framework comprising two key innovations: (1) a coordinate-guided local self-distillation mechanism that relaxes strict spatial alignment constraints between augmented views via the NuLo strategy, enabling cross-scale feature alignment and fine-grained representation learning; and (2) a large-field-of-view semi-supervised fine-tuning strategy to fully exploit unlabeled data. MUSE adopts an encoder-decoder architecture integrating multi-scale feature fusion, self-supervised contrastive learning, and semi-supervised optimization. Evaluated on three major benchmarks, MUSE significantly outperforms existing supervised methods and general-purpose histopathology foundation models, demonstrating superior performance and generalization—particularly in low-label regimes.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Nucleus detection and classification (NDC) in histopathology analysis is a fundamental task that underpins a wide range of high-level pathology applications. However, existing methods heavily rely on labor-intensive nucleus-level annotations and struggle to fully exploit large-scale unlabeled data for learning discriminative nucleus representations. In this work, we propose MUSE (MUlti-scale denSE self-distillation), a novel self-supervised learning method tailored for NDC. At its core is NuLo (Nucleus-based Local self-distillation), a coordinate-guided mechanism that enables flexible local self-distillation based on predicted nucleus positions. By removing the need for strict spatial alignment between augmented views, NuLo allows critical cross-scale alignment, thus unlocking the capacity of models for fine-grained nucleus-level representation. To support MUSE, we design a simple yet effective encoder-decoder architecture and a large field-of-view semi-supervised fine-tuning strategy that together maximize the value of unlabeled pathology images. Extensive experiments on three widely used benchmarks demonstrate that MUSE effectively addresses the core challenges of histopathological NDC. The resulting models not only surpass state-of-the-art supervised baselines but also outperform generic pathology foundation models.

Problem

• Detecting and classifying nuclei in histopathology without dense annotations
• Leveraging unlabeled data for discriminative nucleus representation learning
• Enabling cross-scale alignment through flexible self-distillation mechanisms

Innovation

• Multi-scale dense self-distillation for nucleus detection
• Coordinate-guided local self-distillation without spatial alignment
• Large field-of-view semi-supervised fine-tuning strategy