arXiv 2503.099632025-12-13eess.IVcs.CV

无需参考的脑解剖切片3D重建 via 学习的图谱坐标

Reference-Free 3D Reconstruction of Brain Dissection Slabs via Learned Atlas Coordinates

Martinos Center for Biomedical Imaging (Martinos Center) at Massachusetts General Hospital (MGH) & Harvard Medical School (HMS)（马萨诸塞州总医院（MGH）及哈佛医学院（HMS）的生物医学成像中心（Martinos Center））；Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT)（麻省理工学院（MIT）的计算机科学与人工智能实验室（CSAIL））；Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital-Amager and Hvidovre, Copenhagen, Denmark（丹麦磁共振研究中心、功能与诊断成像及研究中心，哥本哈根大学医院-阿迈厄斯和赫维多尔，哥本哈根，丹麦）；Massachusetts Alzheimer’s Disease Research Center at MGH & HMS（马萨诸塞州总医院（MGH）及哈佛医学院（HMS）的阿尔茨海默病研究中心）；University of Washington（华盛顿大学）；Pathology Department at MGH & HMS（马萨诸塞州总医院（MGH）及哈佛医学院（HMS）的病理部门）；Neurology Department at MGH & HMS（马萨诸塞州总医院（MGH）及哈佛医学院（HMS）的神经病学部门）；Department of（部门）

Lin Tian, Jonathan Williams-Ramirez, Dina Zemlyanker, Lucas J. Deden-Binder, Rogeny Herisse, Theresa R. Connors, Mark Montine, Istvan N Huszar, Lilla Zöllei, Se… 展开作者

Lin Tian, Jonathan Williams-Ramirez, Dina Zemlyanker, Lucas J. Deden-Binder, Rogeny Herisse, Theresa R. Connors, Mark Montine, Istvan N Huszar, Lilla Zöllei, Sean I. Young, Christine Mac Donald, C. Dirk Keene, Derek H. Oakley, Bradley T. Hyman, Oula Puonti, Matthew S. Rosen, Juan Eugenio Iglesias

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

RefFree通过学习图谱坐标实现无需参考的脑切片3D重建，适用于单个切片或部分堆栈，提升重建速度和准确性。

摘要：

神经病理学与MRI之间的相关性有潜力将微观病理特征转移到活体扫描中。越来越多的兴趣在于从切片照片的3D重建堆栈中建立这些相关性，这些照片通常在脑银行解剖过程中常规拍摄。这些照片避免了需要体外MRI，这在广泛普及方面并不普遍。然而，现有方法要么需要一个对应的3D参考（例如，体外MRI扫描或使用结构光扫描器获取的脑表面），要么需要完整的脑切片堆栈，这严重限制了应用范围。在这里，我们提出了RefFree，一种用于解剖照片的3D重建方法，无需外部参考。RefFree利用预测的3D坐标在标准图谱空间（MNI）中作为指导，对任意一组切片（包括单个切片）进行一致的3D体积重建。为了支持RefFree的流程，我们训练了一个图谱坐标预测网络，该网络从2D照片估计坐标图，使用从数字切片3D MRI数据生成的合成照片，通过随机外观增强泛化能力。作为副产品，RefFree可以在没有重建的情况下，将信息（例如，解剖标签）从图谱空间传播到单张照片。在模拟和真实数据上的实验表明，当所有切片都可用时，RefFree的性能与现有经典方法相当，但速度显著提高。此外，RefFree在部分堆栈或甚至单个切片上实现了准确的重建和配准。我们的代码可在https://github.com/lintian-a/reffree上获得。

英文摘要：

Correlation of neuropathology with MRI has the potential to transfer microscopic signatures of pathology to in vivo scans. There is increasing interest in building these correlations from 3D reconstructed stacks of slab photographs, which are routinely taken during dissection at brain banks. These photographs bypass the need for ex vivo MRI, which is not widely accessible. However, existing methods either require a corresponding 3D reference (e.g., an ex vivo MRI scans, or a brain surface acquired with a structured light scanner) or a full stack of brain slabs, which severely limits applicability. Here we propose RefFree, a 3D reconstruction method for dissection photographs that does not require an external reference. RefFree coherently reconstructs a 3D volume for an arbitrary set of slabs (including a single slab) using predicted 3D coordinates in the standard atlas space (MNI) as guidance. To support RefFree's pipeline, we train an atlas coordinate prediction network that estimates the coordinate map from a 2D photograph, using synthetic photographs generated from digitally sliced 3D MRI data with randomized appearance for enhanced generalization. As a by-product, RefFree can propagate information (e.g., anatomical labels) from atlas space to one single photograph even without reconstruction. Experiments on simulated and real data show that, when all slabs are available, RefFree achieves performance comparable to existing classical methods but at substantially higher speed. Moreover, RefFree yields accurate reconstruction and registration for partial stacks or even a single slab. Our code is available at https://github.com/lintian-a/reffree.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。