arXiv cs.AI· Agamdeep S. Chopra, Caitlin Neher, Tianyi Ren, Juampablo E. Heras Rivera, Hesamoddin Jahanian, Mehmet Kurt·· 3 小时前AI 评分28
SFL-Net：用于多对比度 MRI 到 tau-PET 合成的依赖正则化潜通路与成分归因
SFL-Net: Dependence-Regularized Latent Pathways for Multi-Contrast MRI-to-tau-PET Synthesis and Component Attribution
AI 导读
研究者提出 SFL-Net，一种包含共享、T1 特异、FLAIR 特异和联合输入潜通路的向量量化架构，用于从结构 MRI 合成 tau-PET，并借助依赖正则化区分各通路。
正文 · 原文
Abstract:Structural MRI-to-tau-PET synthesis estimates molecular imaging patterns from indirect anatomical correlates. We present SFL-Net, a vector-quantized architecture with shared, T1-specific, FLAIR-specific, and joint-input latent pathways. Dependence regularization differentiates these pathways, while latent-derived structural conditioning supports decoding without direct encoder-decoder feature skips. We evaluated SFL-Net and capacity-matched VAE, VQ-VAE, U-Net, and SPADE-U-Net variants using 605 training and 83 internal validation subjects pooled from ADNI-3 and OASIS-3. SFL-Net achieved the lowest raw PET MAE and MSE and the highest PSNR among the implemented models, while SPADE-U-Net variants performed better on several SUVR and staging metrics. Exact coalitional Shapley attribution assigned the largest average marginal contribution to the joint-input pathway, with substantial variation across subjects. These contributions describe decoder dependence under latent replacement and retained structural conditioning; they do not establish identifiable information-theoretic synergy or the incremental benefit of combining MRI contrasts. Predictions showed dynamic-range compression at higher tau burden, and several staging metrics were close to a constant Stage 1 reference. SFL-Net provides a framework for structured latent-component interrogation in multi-input image synthesis, linking architectural pathways to reconstruction behavior.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| ACM classes: | I.2.10; I.4.8 |
| Cite as: | arXiv:2602.22545 [cs.CV] |
| (or arXiv:2602.22545v4 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2602.22545 arXiv-issued DOI via DataCite |
Submission history
From: Agamdeep Chopra [view email]
[v1]
Thu, 26 Feb 2026 02:37:38 UTC (20,260 KB)
[v2]
Wed, 8 Apr 2026 22:09:04 UTC (5,291 KB)
[v3]
Fri, 3 Jul 2026 00:03:22 UTC (2,203 KB)
[v4]
Wed, 11 Nov 2026 23:57:38 UTC (6,648 KB)
来源：arXiv cs.AI · arxiv.org