Wu Xue, Eggebrecht Adam T, Ferradal Silvina L, Culver Joseph P, Dehghani Hamid
School of Computer Science, University of Birmingham, Birmingham, B15 2TT, UK.
Department of Radiology, Washington University School of Medicine, 4525 Scott Avenue, St. Louis, MO, 63110, USA.
Biomed Opt Express. 2015 Oct 26;6(11):4567-84. doi: 10.1364/BOE.6.004567. eCollection 2015 Nov 1.
Real-time imaging of human brain has become an important technique within neuroimaging. In this study, a fast and efficient sensitivity map generation based on Finite Element Models (FEM) is developed which utilises a reduced sensitivitys matrix taking advantage of sparsity and parallelisation processes. Time and memory efficiency of these processes are evaluated and compared with conventional method showing that for a range of mesh densities from 50000 to 320000 nodes, the required memory is reduced over tenfold and computational time fourfold allowing for near real-time image recovery.
人脑的实时成像已成为神经成像中的一项重要技术。在本研究中，开发了一种基于有限元模型（FEM）的快速高效的灵敏度图生成方法，该方法利用稀疏性和并行化过程的简化灵敏度矩阵。对这些过程的时间和内存效率进行了评估，并与传统方法进行了比较，结果表明，对于从50000到320000个节点的一系列网格密度，所需内存减少了十倍以上，计算时间减少了四倍，从而实现了近实时图像恢复。
