CISE Research Instrumentation: Computational Techniques for 3-D Modeling and Analysis of Infrared and MRI Image Sequences

Goutsias, John I. Prince, Jerry L.约翰霍普金斯大学CISE研究仪器：用于红外和MRI图像序列三维建模和分析的计算技术。该研究仪器奖用于获得一台高速超级计算机，这将使以下研究项目成为可能：-从光学和红外视频序列中自动检测、识别和跟踪目标的新技术，-从三维磁共振图像中分割、建模和统计表征人脑皮层的几何形状，以及-从标记磁共振图像中重建MR材料标记的样条算法。提议的设备将致力于研究红外和磁共振成像（MRI）数据建模和分析的新型计算技术。所要求的设备将：(a)允许使用3-D形态图像处理和分析技术开发和实施实时和近实时算法（项目1）。(b)允许快速实施连续图像建模和主动几何算法，以将拓扑正确的表面拟合到皮层的中央层；开发有效的方法来统计表征皮层主要沟（裂隙）的形状（项目2）。(c)允许开发和实施用于运动估计的3- d光流算法，并允许求解大型矩阵方程以确定用于标签表面拟合和分析的高阶样条系数（项目3）。

9729576 Goutsias, John I. Prince, Jerry L. Johns Hopkins University CISE Research Instrumentation: Computational Techniques for 3-D Modeling and Analysis of Infrared and MRI Image Sequences This research instrumentation award serves to acquire a high-speed supercomputer which will enable the following research projects: - New Techniques in Automatic Target Detection, Recognition, and Tracking from Optical and Infrared Video Sequences, - Segmentation, Modeling, and Statistical Characterization of the Geometry of the Human Brain Cortex from Three-Dimensional Magnetic Resonance Images, and -Spline Algorithms for the Reconstruction of MR Material Markers from Tagged Magnetic Resonance Images. The proposed equipment will be dedicated to research in novel computational techniques for modeling and analysis of infrared and magnetic resonance imaging (MRI) data. The requested equipment will: (a) Allow development and implementation of real-time and near-real-time algorithms using 3-D morphological image processing and analysis techniques (Project 1). (b) Allow fast implementation of continuous image modeling and active geometry algorithms to fit a topologically correct surface to the central layer of the cortex, and development of efficient methods to statistically characterize the shapes of the major sulci (fissures) of the cortex (Project 2). (c) Allow development and implementation of 3-D optical flow algorithms for motion estimation, and enable solution of large matrix equations to determine higher order spline coefficients for tag surface fitting and analysis (Project 3).