用对比信息增强和对抗训练，让成人MRI模型更好重建新生儿图像。
Contrast-Informed Augmentation and Domain-Adversarial Training for Adult-to-Neonatal MR Reconstruction Generalization
- 引入对比感知数据增强和域对抗训练提升泛化能力
- 在加速因子8下SSIM达0.848，显著优于基线模型
- 适合需要跨年龄群体鲁棒重建的医学影像研究者
目的：探究对比感知数据增强与域对抗训练是否能提升端到端变网（E2E-VarNet）从成人到新生儿MRI重建的泛化性能。方法：比较三种训练策略：（1）仅使用未增强的成人数据训练；（2）混合训练，包含未增强成人数据与对比感知增强的成人数据；（3）混合训练结合域对抗目标。模型在回顾性欠采样的多线圈成人T2加权脑部MRI数据上训练，并在加速因子R=4和R=8下，于新生儿和成人测试集上评估，采用定量指标与定性分析。特征分析考察域对抗训练对未增强成人、增强成人及新生儿测试样本隐空间表示的影响。结果：混合训练（Mixed）与混合域对抗训练（Mixed-DAT）在新生儿数据上均优于仅成人训练（Unaug-Only）。在R=4时，Mixed-DAT表现最佳（SSIM = 0.924 ± 0.027，PSNR = 33.98 ± 1.15 dB）；在R=8时，Mixed-DAT在SSIM上最优（0.848 ± 0.031，优于Unaug-Only的0.766 ± 0.037和Mixed的0.814 ± 0.035），Mixed在PSNR上最优（29.56 ± 0.83 dB，优于Unaug-Only的26.26 ± 0.78 dB和Mixed-DAT的29.43 ± 0.83 dB）。t-SNE图定性分析显示，Mixed-DAT提升了未增强成人、增强成人与新生儿测试样本间隐表示的重叠度。结论：对比感知增强与域对抗训练可有效提升深度学习MRI重建在成人到新生儿迁移中的泛化能力，表明该组合有助于应对欠采样新生儿MRI中的域偏移问题。
原文摘要 · Abstract (English)
Purpose: To investigate whether contrast-informed data augmentation and domain-adversarial training improve the adult-to-neonatal generalization of the E2E-VarNet. Methods: Three training regimes were investigated: (1) adult-only training with unaugmented adult data, (2) mixed training with paired unaugmented and neonatal-informed augmented adult data, and (3) mixed training with a domain-adversarial objective. Models were trained on retrospectively undersampled multi-coil adult T2-weighted brain MR data and evaluated on neonatal and adult test data at acceleration factors $R=4$ and $R=8$ using quantitative metrics and qualitative evaluation. Feature analyses assessed whether domain-adversarial training altered the latent representations of unaugmented adult, augmented adult, and neonatal test samples. Results: Mixed training (Mixed) and mixed domain-adversarial training (Mixed-DAT) outperformed unaugmented adult-only training (Unaug-Only) when evaluated on neonatal data. At R=4, Mixed-DAT achieved the best performance (SSIM = 0.924 +/- 0.027, PSNR = 33.98 +/- 1.15 dB). At R=8, Mixed-DAT performed best when measured using SSIM (0.848 +/- 0.031 vs. 0.766 +/- 0.037 for Unaug-Only and 0.814 +/- 0.035 for Mixed) and Mixed performed best when measured using PSNR (29.56 +/- 0.83 dB vs. 26.26 +/- 0.78 dB for Unaug-Only and 29.43 +/- 0.83 dB for Mixed-DAT). Qualitative assessment of t-SNE plots suggested that Mixed-DAT increased the overlap among the latent representations of the unaugmented adult, augmented adult, and neonatal test data. Conclusion: Contrast-informed augmentation and domain-adversarial training improved adult-to-neonatal generalization of deep learning-based MR reconstruction. These findings suggest that contrast-informed data augmentation combined with adversarial training may improve robustness to domain shift in undersampled neonatal MR reconstruction.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。