Can Score-Based Generative Modeling Effectively Handle Medical Image Classification?

2025-02-24

AI Summary

This work addresses the challenges of data scarcity and limited diversity in medical image classification. We propose the first likelihood-guided classification paradigm that reformulates score-based generative modeling into a supervised classification framework. Unlike conventional discriminative models, our approach alleviates reliance on large-scale labeled datasets while jointly optimizing generative fidelity and discriminative accuracy. Specifically, we apply this method to breast X-ray classification, achieving state-of-the-art performance on three benchmark public datasets—CBIS-DDSM, INbreast, and VinDr-Mammo. The framework demonstrates robust generalization under low-data regimes, offering a principled solution for few-shot medical imaging analysis. To foster reproducibility and further research, we release the implementation publicly. This work establishes a novel paradigm that bridges generative modeling and supervised learning for diagnostic imaging tasks, advancing both methodological rigor and clinical applicability.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Reasoning under Uncertainty: Graphical Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The remarkable success of deep learning in recent years has prompted applications in medical image classification and diagnosis tasks. While classification models have demonstrated robustness in classifying simpler datasets like MNIST or natural images such as ImageNet, this resilience is not consistently observed in complex medical image datasets where data is more scarce and lacks diversity. Moreover, previous findings on natural image datasets have indicated a potential trade-off between data likelihood and classification accuracy. In this study, we explore the use of score-based generative models as classifiers for medical images, specifically mammographic images. Our findings suggest that our proposed generative classifier model not only achieves superior classification results on CBIS-DDSM, INbreast and Vin-Dr Mammo datasets, but also introduces a novel approach to image classification in a broader context. Our code is publicly available at https://github.com/sushmitasarker/sgc_for_medical_image_classification

Problem

• Explore score-based generative models for medical image classification
• Address data scarcity and diversity in medical datasets
• Propose a novel generative classifier for mammographic images

Innovation

• Score-based generative models
• Medical image classification
• Superior classification results