MUSE: Harnessing Precise and Diverse Semantics for Few-Shot Whole Slide Image Classification

TL;DR

MUSE introduces a novel framework for few-shot whole slide image classification that enhances semantic precision and diversity through sample-wise adaptation and retrieval-augmented multi-view generation, improving generalization in limited supervision scenarios.

Contribution

The paper proposes MUSE, a new method combining sample-wise semantic refinement and multi-view generation to improve few-shot pathology image classification.

Findings

Outperforms existing vision-language baselines on three benchmark datasets.

Effective semantic optimization enhances model robustness and reduces overfitting.

Sample-aware semantic refinement improves generalization in limited data settings.

Abstract

In computational pathology, few-shot whole slide image classification is primarily driven by the extreme scarcity of expert-labeled slides. Recent vision-language methods incorporate textual semantics generated by large language models, but treat these descriptions as static class-level priors that are shared across all samples and lack sample-wise refinement. This limits both the diversity and precision of visual-semantic alignment, hindering generalization under limited supervision. To overcome this, we propose the stochastic MUlti-view Semantic Enhancement (MUSE), a framework that first refines semantic precision via sample-wise adaptation and then enhances semantic richness through retrieval-augmented multi-view generation. Specifically, MUSE introduces Sample-wise Fine-grained Semantic Enhancement (SFSE), which yields a fine-grained semantic prior for each sample through MoE-based adaptive visual-semantic interaction. Guided by this prior, Stochastic Multi-view Model Optimization (SMMO) constructs an LLM-generated knowledge base of diverse pathological descriptions per class, then retrieves and stochastically integrates multiple matched textual views during training. These dynamically selected texts serve as enriched semantic supervisions to stochastically optimize the vision-language model, promoting robustness and mitigating overfitting. Experiments on three benchmark WSI datasets show that MUSE consistently outperforms existing vision-language baselines in few-shot settings, demonstrating that effective few-shot pathology learning requires not only richer semantic sources but also their active and sample-aware semantic optimization. Our code is available at: https://github.com/JiahaoXu-god/CVPR2026_MUSE.