发表机构
Yonsei University; Microsoft Research Asia; Korea Institute of Science and Technology（延世大学; 微软亚洲研究院; 韩国科学技术研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对多对比度MRI合成，提出统一多平面自回归扩散（MPAD），利用二维扩散和平面间先验实现高效三维体积生成，显著降低计算成本并支持一对多翻译。
AI 中文摘要
获取完整的磁共振成像（MRI）对比度序列耗时且让患者感到不适，尽管多对比度成像具有诊断价值。这促使从已获取的对比度中合成缺失的对比度，这本质上是一个三维问题，需要轴向、矢状面和冠状面之间的解剖学一致性。然而，完全三维的生成模型在计算资源随体积大小立方级增长的情况下往往不切实际。我们提出了一种统一的多平面自回归扩散（MPAD），这是一种潜在扩散框架，通过高效的逐平面二维操作实现全体积三维合成，同时保持体积一致性。一个三维自编码器首先将MRI扫描压缩为各向同性的三维潜在表示。然后训练一个二维扩散模型，以源对比度切片和未掩蔽的目标对比度切片为条件，重建目标对比度的掩蔽潜在切片。在推理过程中，我们引入了带有平面间先验的逐平面自回归合成。切片在一个平面方向内以随机顺序自回归生成，以保持平面内连续性，然后作为条件先验传播到正交平面方向，以强制平面间一致性。与三维潜在扩散基线相比，MPAD将训练和推理的FLOPs分别减少了7倍和3倍，同时降低了推理时间和峰值内存消耗。在多个数据集上的实验表明，MPAD实现了优越的性能，生成高保真度的三维体积，并在单个统一模型中支持一对多翻译。
英文摘要
Acquiring a complete set of magnetic resonance imaging (MRI) contrasts is time-intensive and uncomfortable for patients, despite the diagnostic value of multi-contrast imaging. This motivates synthesizing missing contrasts from those already acquired, which is an inherently 3D problem requiring anatomical coherence across axial, sagittal, and coronal planes. However, fully 3D generative models are often impracti- cal under computational resources that scale cubically with volume size. We propose a unified Multi-Plane Autoregressive Diffusion (MPAD), a latent diffusion framework that achieves full-volume 3D synthesis using efficient plane-wise 2D operations while preserving volumetric coherence. A 3D autoencoder first compresses MRI scans into an isotropic 3D la- tent representation. A 2D diffusion model is then trained to reconstruct masked latent slices of the target contrast, conditioned on both source- contrast slices and unmasked target-contrast slices. During inference, we introduce plane-wise autoregressive synthesis with inter-plane priors. Slices are generated autoregressively in random order within one plane orientation to maintain intra-plane continuity, then propagated as con- ditioning priors to orthogonal plane orientations to enforce inter-plane consistency. Compared to 3D latent diffusion baselines, MPAD reduces training and inference FLOPs by 7x and 3x, respectively, while also lowering inference time and peak memory consumption. Experiments on multiple datasets demonstrate that MPAD achieves superior perfor- mance, generating high-fidelity 3D volumes and supporting one-to-many translation within a single unified model.
CommentsAccepted to ECCV 2026
Journal refECCV 2026. Lecture Notes in Computer Science, vol 17034. Springer, Cham
DOI：10.1007/978-3-032-37450-9_16
