Erem Burak, Brooks Dana H
Center for Communications and Digital Signal Processing, Dept. of Electrical & Computer Engineering Northeastern University, 360 Huntington Ave., Boston, MA.
Proc IEEE Int Symp Biomed Imaging. 2011 Jun 9;2011(March 30 2011-April 2 2011):504-507. doi: 10.1109/ISBI.2011.5872455.
In a number of medical imaging modalities, including measurements or estimates of electrical activity on cortical or cardiac surfaces, it is often useful to estimate spatial derivatives of data on curved anatomical surfaces represented by triangulated meshes. Assuming the triangle vertices are points on a smooth manifold, we derive a method for estimating gradients and Hessians on locally 2D surfaces embedded in 3D directly in the global coordinate system. Accuracy of the method is validated through simulations on both smooth and corrugated surfaces.
在许多医学成像模态中，包括对皮质或心脏表面电活动的测量或估计，估计由三角网格表示的弯曲解剖表面上数据的空间导数通常很有用。假设三角形顶点是光滑流形上的点，我们推导了一种直接在全局坐标系中估计嵌入在3D中的局部2D表面上的梯度和黑塞矩阵的方法。该方法的准确性通过在光滑和波纹表面上的模拟得到验证。
