用普通MRI生成脂肪分数图，轻量U-Net比扩散模型快200倍还更准。
Do We Really Need Diffusion? A Fast U-Net for Paired Medical Image Translation
- 用4层轻量U-Net实现T2w到SFF的图像翻译。
- 相关系数达0.975，误差仅0.014，优于扩散模型。
- 推理速度提升208倍，适合临床实时应用。
磁共振成像信号脂肪分数（MRI-SFF）是代谢和骨骼肌肉疾病的重要生物标志物，但其获取需特殊序列，难以常规使用。本文研究是否可从广泛可用的T2加权（T2w）MRI中通过图像到图像翻译（I2I）估计SFF。在德国国家队列（NAKO）数据集上，使用230,048张配对2D图像（训练183,517张，验证23,621张，测试22,910张），对比了轻量级4层U-Net与最先进的去噪扩散概率模型（DDPM）。两者均显著优于恒等映射基线（皮尔逊相关系数r=0.769，平均绝对误差MAE=0.070±0.054），证实模型学习到了非平凡的跨模态映射。有趣的是，轻量U-Net在相关性（r=0.975 vs. 0.962）和误差（MAE=0.014±0.015 vs. 0.019±0.019）上均优于DDPM，且推理时间减少208倍（每张图像25.2毫秒 vs. 5,227.2毫秒，使用50步去噪扩散隐式模型DDIM）。优异的临床表现与极低计算成本使其具备实时临床应用潜力。
原文摘要 · Abstract (English)
Magnetic resonance imaging-signal fat fraction (MRI-SFF) quantifies tissue fat and serves as an established biomarker for metabolic and musculoskeletal disorders. The acquisition requires, however, specialized MRI sequences, which are not available routinely. We investigate whether SFF can be estimated from widely available T2-weighted (T2w) MRI via image-to-image translation (I2I). We further compare a lightweight 4-level U-Net to a state-of-the-art Denoising Diffusion Probabilistic Model (DDPM) using a dataset of 230 048 paired 2D images (183 517 train, 23 621 val, 22 910 test) from the German National Cohort (NAKO). Both models clearly outperform the identity baseline (Pearson correlation r = 0.769, mean absolute error MAE = 0.070 +/- 0.054), which confirms that the models learn a non-trivial cross-modal mapping. Interestingly, the lightweight U-Net outperforms the DDPM in both correlation (r = 0.975 vs. 0.962) and error (MAE = 0.014 +/- 0.015 vs. 0.019 +/- 0.019), while reducing inference time by a factor of 208 (25.2 ms vs. 5 227.2 ms per image using 50 Denoising Diffusion Implicit Model (DDIM) steps). The strong clinical performance at substantially reduced computational cost enables real-time clinical use.
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。