SBIR Phase II: A 3-D Image Matrix Processor

9628968 Browne这个小型企业创新研究第二阶段项目是ARACOR计划的关键部分，该计划旨在开发用于三维(3-D)成像应用的实用体积计算机断层扫描(CT)扫描仪。在第一阶段计划中，开发了一种专门的硬件加速器--图像矩阵处理器(IMP)的高级设计。IMP能够使用主要的重建算法执行低成本、高速的3D重建：VOVAL、Grangeat和Feldkamp。IMP真正的创新之处在于，它能够在单一设计下同时进行2-D和3-D重建。这是一项重大的技术创新，消除了商业3D CT发展的主要技术障碍：处理体数据所需的极长的重建时间。目标是通过展示基于IMP的工作站的实用性，在第一阶段取得成功的基础上再接再厉，其中，实用性是以速度和负担能力为标准的。特别是，第二阶段研究的技术目标是完善IMP设计；通过详细、门级仿真定制集成电路(IC)来评估其性能；生产和测试足够的IC以构建一个IMP电路板；以及表征和基准具有约2小时重建10243图像的能力的工作站(以今天的中央处理器速度)可以比具有类似性能的多处理器工作站的构建成本低一个数量级。这一能力相当的工作站将首次使需要创建和可视化大体积图像的新兴医疗、工业和科学应用程序在普通桌面环境的范围内成为可能。IMP的商业利益可能被证明对美国的工业、医疗和科学市场极其重要。实用型工业CT系统的出现将开启制造业的新纪元，它将体现诸如逆向工程、快速原型制造和实体自由形状制造等新兴概念。在医学和科学界，高分辨率三维CT仪器将成为骨质疏松症研究的体外研究、肿瘤学和其他研究的小型实验室动物体内研究以及手术规划和三维建模等新型三维可视化技术不可缺少的工具。

9628968 Browne This Small Business Innovation Research Phase II project is a key part of an ARACOR initiative to develop a practical volumetric computed tomography (CT) scanner for three-dimensional (3-D) imaging applications. In the Phase I program, a high-level design for a specialized hardware accelerator, an Image Matrix Processor (IMP), was developed. The IMP is capable of performing, low-cost, high-speed 3-D reconstruction's with the principal reconstruction algorithms: VOIR, Grangeat, and Feldkamp. The truly innovative aspect of the IMP is its ability to perform both 2-D and 3-D reconstruction's with a single design. This is a significant technological innovation, eliminating the major technical obstacle to the progress of commercial 3-D CT: the exceedingly long reconstruction times required to process volumetric data. The goal is to build upon the successful Phase I effort by demonstrating the practicality of an IMP-based workstation, where practicality is defined in terms of speed and affordability. In particular, the technical objectives of the Phase II study are to refine the IMP design; assess its performance through detailed, gate-level simulations of the custom integrated circuits (ICs) which lie at the heart of the IMP; produce and test enough ICs to build one IMP circuit board; and characterize and benchmark the workstation with the power to reconstruct 10243 images in about an 2 hours (at today's CPU speeds) can be constructed for an order of magnitude less cost than a multiprocessor workstation capable of similar performance. This workstation of comparable power, will, for the first time, make it feasible for emerging medical, industrial and scientific applications requiring the creation and visualization of large volumetric images, to be within the domain of an ordinary desk-top environment. The commercial benefit of the IMP could prove to be extremely significant for America's industrial, medical, and scientific markets. The advent of practical volumetric indust rial CT systems will usher in a new era in manufacturing that will embody emerging concepts such as reverse engineering, rapid prototyping, and solid freeform fabrication. In the medical and scientific communities, and high resolution 3-D CT instrument will be an indispensable tool for in-vitro study of bone samples for osteoporosis research, for in-vivo study of small laboratory animals for oncology and other research, and novel 3-D visualization technologies such as surgical planning and 3-D modeling.