Beyond a Single Mode: GAN Ensembles for Diverse Medical Data Generation

2025-03-31

AI Summary

Generative AI for medical image synthesis faces a trilemma among fidelity, diversity, and computational efficiency; moreover, single GAN models often suffer from mode collapse and inadequate data distribution coverage. To address this, we propose a multi-objective optimization framework for GAN ensembles tailored to comprehensive coverage of medical data distributions. Leveraging the NSGA-II algorithm, our method dynamically selects complementary, low-redundancy heterogeneous subnetworks—spanning 22 distinct architectures, diverse loss functions, and regularization strategies—thereby overcoming inherent limitations of individual models. The framework preserves computational efficiency while significantly improving generation quality: FID decreases by 18.7%, LPIPS-based diversity increases by 32.4%, and downstream diagnostic model performance is enhanced. To the best of our knowledge, this is the first systematic multi-objective optimization approach specifically designed for GAN ensemble selection in medical imaging.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Ensemble Methods
• Natural Language Processing: Generation

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

The advancement of generative AI, particularly in medical imaging, confronts the trilemma of ensuring high fidelity, diversity, and efficiency in synthetic data generation. While Generative Adversarial Networks (GANs) have shown promise across various applications, they still face challenges like mode collapse and insufficient coverage of real data distributions. This work explores the use of GAN ensembles to overcome these limitations, specifically in the context of medical imaging. By solving a multi-objective optimisation problem that balances fidelity and diversity, we propose a method for selecting an optimal ensemble of GANs tailored for medical data. The selected ensemble is capable of generating diverse synthetic medical images that are representative of true data distributions and computationally efficient. Each model in the ensemble brings a unique contribution, ensuring minimal redundancy. We conducted a comprehensive evaluation using three distinct medical datasets, testing 22 different GAN architectures with various loss functions and regularisation techniques. By sampling models at different training epochs, we crafted 110 unique configurations. The results highlight the capability of GAN ensembles to enhance the quality and utility of synthetic medical images, thereby improving the efficacy of downstream tasks such as diagnostic modelling.

Problem

• Overcoming mode collapse in GANs for medical imaging
• Balancing fidelity and diversity in synthetic data generation
• Optimizing GAN ensembles for efficient medical image synthesis

Innovation

• GAN ensembles for diverse medical data
• Multi-objective optimization balances fidelity diversity
• Optimal GAN selection enhances synthetic images