Dual-Domain Equivariant Generative Adversarial Network for Multimodal CT-PET Synthesis

2026-06-11

AI Summary

Current generative adversarial networks for CT-PET multimodal image synthesis rely solely on spatial-domain information and neglect geometric consistency, resulting in insufficient anatomical fidelity. To address this limitation, this work proposes DDE-GAN, which uniquely integrates dual-domain learning with a physics-driven rotation-equivariant constraint. By jointly modeling both spatial and frequency domains and employing a hierarchical dual-domain training strategy, the method incorporates a multi-stage loss function to enforce intra- and inter-domain consistency. Evaluated on the HECKTOR 2022 dataset, DDE-GAN significantly outperforms existing baselines, yielding synthetic images with enhanced anatomical accuracy and robustness, thereby facilitating clinical applications such as PET image completion and data augmentation.

Abstract

We present a Dual-Domain Equivariant Generative Adversarial Network (DDE-GAN) for multimodal CT-PET image synthesis. Traditional GAN-based approaches often operate solely in the spatial domain and ignore geometric consistency, resulting in limited structural fidelity. DDE-GAN addresses these challenges by jointly learning from both spatial and frequency (Fourier) domains, capturing complementary anatomical and spectral information. Furthermore, rotational equivariance embedded in the physics of the CT and PET measurements are integrated into the loss of both the generator and discriminator to ensure consistent responses under rotations, improving anatomical accuracy. A hierarchical dual-domain training strategy enforces intra- and inter-domain consistency through multi-stage loss functions. Evaluated on the HECKTOR 2022 CT-PET dataset, DDE-GAN achieves superior synthesis quality over baseline models for CT-PET image synthesis. The results demonstrate that combining dual-domain learning with geometric equivariance substantially enhances multimodal image synthesis accuracy and robustness, enabling practical applications in PET completion and data augmentation.

Problem

• multimodal image synthesis
• CT-PET synthesis
• structural fidelity
• geometric consistency
• spatial domain

Innovation

• dual-domain learning
• rotational equivariance
• multimodal image synthesis
• CT-PET synthesis
• generative adversarial network