A FPGA-based Real-Time Volume Rendering System

描述（由申请人提供）：这个小企业创新研究第二阶段项目涉及一个基于FPGA的实时体绘制系统的研究和开发。医学放射科医生面临着数据过载的危机，这种危机每年都在恶化。这一危机是由诊断成像技术（即CT、MR和超声）的不断进步造成的，这些技术每年都能产生更多分辨率更高的图像。使这个问题更加复杂的是，成像研究的数量每年也在增长。当前的硬件和软件解决方案无法应对挑战。因此，重要的数据往往被忽视，产生不准确的医学诊断。该项目的目标是开发一种低成本的基于FPGA的体积渲染，以缓解这种数据过载危机。通过消除对3D图像没有贡献的数据，体积渲染器的性能将远远超过当前的解决方案。通过基于FPGA，体积渲染器将是可移植的，允许未来的性能和功能增强。此外，渲染器将通过使用高阶渐变来支持非常高质量的渲染。因此，体积渲染器将吸引许多市场的用户，包括：医疗诊断成像，石油勘探，安全检查和工程模拟。体积绘制器的4D功能的一个非常令人兴奋的应用将是多层CT心脏成像，使得能够执行非侵入性程序来代替当前的侵入性程序。因此，国家将受益于生产力的提高和医疗保健的改善。该项目中开发的体积渲染器将能够实时查看由CT和MR扫描仪生成的非常大的3D和4D数据集。这将使医学成像更加准确和有用，改善放射工作流程，加快药物的发现。此外，4D心脏渲染等新的高端应用也将成为可能。

DESCRIPTION (provided by applicant): This Small Business Innovation Research Phase II project involves the research and development of a real-time FPGA-based volume rendering system. Medical radiologists face a data overload crisis that is getting worse with each year. This crisis is caused by ongoing advances in diagnostic imaging technologies (i.e. CT, MR and ultrasound) that produce more images with higher resolution each year. Compounding this problem is the fact that the number of imaging studies is also growing each year. Current hardware and software solutions are not up to the challenge. As a result, important data is often overlooked producing inaccurate medial diagnoses. The goal of this project is to develop a low-cost FPGA-based volume render that will alleviate this data overload crisis. By eliminating data that does not contribute to 3D images, the volume Tenderer's performance will far exceed current solutions. By being based on FPGAs, the volume renderer will be upgradeable, allowing for future performance and functionality enhancements. Additionally, the renderer will support very high quality renderings through the use of higher-order gradients. Thus, the volume renderer will appeal to users in many markets, including: medical diagnostic imaging, oil exploration, security inspection and engineering simulations. One very exciting application of the 4D capabilities of the volume renderer will be in multi-slice CT cardiac imaging, enabling non-invasive procedures to be performed in place of current invasive procedures. As a consequence, the nation stands to benefit from productivity increases and improved health care. The volume renderer developed in this project will enable very large 3D and 4D data sets, generated by CT and MR scanners, to be viewed in real-time. This will make medical imaging more accurate and useful, improve the radiological workflow and speed up the discover of drugs. In addition, new high-end application such as 4D cardiac rendering will become possible.