Rethinking Bias in Generative Data Augmentation for Medical AI: a Frequency Recalibration Method

2025-11-15

AI Summary

This work addresses performance degradation in medical generative data augmentation (GDA) caused by frequency-domain distribution mismatch—termed *frequency bias*—between real and synthetic images. We propose FreRec, a training-free post-processing framework that identifies frequency misalignment as the primary cause of synthetic image quality degradation. FreRec introduces a two-stage frequency-domain calibration strategy: Statistical High-frequency Replacement (SHR) and Reconstructed High-frequency Mapping (RHM), which jointly recalibrate and reconstruct high-frequency components in synthetic images. The method is model-agnostic and seamlessly integrates into existing GDA pipelines. Extensive evaluation across diverse medical imaging modalities—including brain MRI, chest X-ray, and fundus photography—demonstrates that FreRec consistently improves classification accuracy and effectively mitigates performance collapse under low-data regimes. By bridging the frequency-domain gap between real and synthetic data, FreRec establishes a novel paradigm for trustworthy medical AI data augmentation.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Developing Medical AI relies on large datasets and easily suffers from data scarcity. Generative data augmentation (GDA) using AI generative models offers a solution to synthesize realistic medical images. However, the bias in GDA is often underestimated in medical domains, with concerns about the risk of introducing detrimental features generated by AI and harming downstream tasks. This paper identifies the frequency misalignment between real and synthesized images as one of the key factors underlying unreliable GDA and proposes the Frequency Recalibration (FreRec) method to reduce the frequency distributional discrepancy and thus improve GDA. FreRec involves (1) Statistical High-frequency Replacement (SHR) to roughly align high-frequency components and (2) Reconstructive High-frequency Mapping (RHM) to enhance image quality and reconstruct high-frequency details. Extensive experiments were conducted in various medical datasets, including brain MRIs, chest X-rays, and fundus images. The results show that FreRec significantly improves downstream medical image classification performance compared to uncalibrated AI-synthesized samples. FreRec is a standalone post-processing step that is compatible with any generative model and can integrate seamlessly with common medical GDA pipelines.

Problem

• Addressing frequency misalignment between real and synthetic medical images
• Reducing harmful biases in generative data augmentation for medical AI
• Improving downstream medical classification through frequency distribution calibration

Innovation

• Frequency Recalibration method reduces distributional discrepancy
• Statistical High-frequency Replacement aligns frequency components
• Reconstructive High-frequency Mapping enhances image details