Abstract:Generative adversarial network (GAN) has been widely explored for cross-modality medical image synthesis. The existing GAN models usually adversarially learn a global sample space mapping from the source-modality to the target-modality and then indiscriminately apply this mapping to all samples in the whole space for prediction. However, due to the scarcity of training samples in contrast to the complicated nature of medical image synthesis, learning a single global sample space mapping that is “optimal” to all samples is very challenging, if not intractable. To address this issue, this paper proposes sample-adaptive GAN models, which not only cater for the global sample space mapping between the source- and the target-modalities but also explore the local space around each given sample to extract its unique characteristic. Specifically, the proposed sample-adaptive GANs decompose the entire learning model into two cooperative paths. The baseline path learns a common GAN model by fitting all the training samples as usual for the global sample space mapping. The new sample-adaptive path additionally models each sample by learning its relationship with its neighboring training samples and using the target-modality features of these training samples as auxiliary information for synthesis. Enhanced by this sample-adaptive path, the proposed sample-adaptive GANs are able to flexibly adjust themselves to different samples, and therefore optimize the synthesis performance. Our models have been verified on three cross-modality MR image synthesis tasks from two public datasets, and they significantly outperform the state-of-the-art methods in comparison. Moreover, the experiment also indicates that our sample-adaptive strategy could be utilized to improve various backbone GAN models. It complements the existing GANs models and can be readily integrated when needed.

Cross-Model Transformer Method for Medical Image Synthesis

Data-Enabled Intelligence in Complex Industrial Systems Cross-Model Transformer Method for Medical Image Synthesis

OA-GAN: Organ-Aware Generative Adversarial Network for Synthesizing Contrast-Enhanced Medical Images

Towards cross-modal organ translation and segmentation: A cycle- and shape-consistent generative adversarial network

TCGAN: a transformer-enhanced GAN for PET synthetic CT

Optimization of Cardiac Magnetic Resonance Synthetic Image Based on Simulated Generative Adversarial Network

SynthMix: Mixing up Aligned Synthesis for Medical Cross-Modality Domain Adaptation

Cross Attention Multi Scale CNN-Transformer Hybrid Encoder is General Medical Image Learner.

An Attentive-based Generative Model for Medical Image Synthesis

Multi-Constraint Transferable Generative Adversarial Networks for Cross-Modal Brain Image Synthesis

Bi-Modality Medical Image Synthesis Using Semi-Supervised Sequential Generative Adversarial Networks

Enhancing Medical Imaging with GANs Synthesizing Realistic Images from Limited Data

Medical Image Synthesis with Generative Adversarial Networks for Tissue Recognition

Sample-adaptive GANs: Linking Global and Local Mappings for Cross-modality MR Image Synthesis.

Robust Cross-modal Medical Image Translation Via Diffusion Model and Knowledge Distillation

FedMed-GAN: Federated Domain Translation on Unsupervised Cross-Modality Brain Image Synthesis

ReeGAN: MRI image edge-preserving synthesis based on GANs trained with misaligned data

Application and prospect for generative adversarial networks in cross-modality reconstruction of medical images.

Swin transformer-based GAN for multi-modal medical image translation

HLSNC-GAN: Medical Image Synthesis Using Hinge Loss and Switchable Normalization in CycleGAN

Multimodal MR Image Synthesis Using Gradient Prior and Adversarial Learning