Computational techniques for processing and visualising very large datasets in 3D

用于处理和可视化 3D 大型数据集的计算技术

• 批准号: 261435-2007
• 负责人: Mudur, Sudhir
• 金额: $ 1.89万
• 依托单位: Concordia University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2009
• 资助国家: 加拿大
• 起止时间: 2009-01-01 至 2010-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=426606
• 关键词: Computational; techniques; processing; visualising; very

The long term objective of my research programme is to be able to create computing environments in which humans can intuitively interact, comprehend complex behaviour and relationships more easily through suitable visualisations, and productively collaborate to accomplish real worlds tasks efficiently. Rapid advances in sensor based data acquisition, real time data logging and voluminous data from complex simulations have led to the situation that data sizes grow much faster than advances in graphics hardware acceleration or in algorithmic techniques. Most 3D processing, rendering and visualization techniques do not scale to these data sizes, certainly not with standard commodity hardware. New computational techniques that can efficiently process and enable 3D visualization of very large and complex data sets are essential. The proposed research directly builds on my previous work during my tenure in India (until 2002), and more specifically on the research programme and collaborations initiated after joining Concordia in 2002. With funding support from NSERC, CFI and other agencies, over the last 4 years I have set up the "Advanced 3D Graphics and Visual Computing Laboratory" at Concordia. This lab is currently equipped with a two-head 3D scanner, virtual and augmented reality devices, a powerful 3D graphics cluster and a suite of high end graphics workstations. Specifically, we shall investigate: 1)Computational techniques for closely coupling visualisation with macro-level feature/pattern discovery techniques applicable to large data sets; this forms the core of this research, 2)New methods of processing raw 3D scan data into a format readily useful in applications such as gaming, cinema, engineering, security, etc. 3)Functionality and data distribution on CPUs and GPUs (looked upon as co-processors) for very large datasets using the graphics cluster in the lab. and 4)Development of interactive and immersive systems incorporating these techniques for multidisciplinary applications in web usage, bioinformatics, patient education and clinical data. The results would immensely benefit the graphics research community at large as well as Canadian industries of gaming, cinema, healthcare and engineering.

我的研究计划的长期目标是能够创建计算环境，其中人类可以直观地进行交互，通过适当的可视化更容易地理解复杂的行为和关系，并有效地合作，以有效地完成真实的世界任务。基于传感器的数据采集、真实的时间数据记录和来自复杂模拟的大量数据的快速发展已经导致数据大小比图形硬件加速或算法技术的发展增长得快得多的情况。大多数3D处理、渲染和可视化技术都无法扩展到这些数据大小，当然也无法使用标准的商用硬件。能够有效处理非常大且复杂的数据集并实现其3D可视化的新计算技术至关重要。拟议的研究直接建立在我在印度任职期间（直到2002年）的工作基础上，更具体地说，建立在2002年加入Concordia后发起的研究计划和合作基础上。在NSERC，CFI和其他机构的资助下，在过去的4年里，我在Concordia建立了“高级3D图形和视觉计算实验室”。该实验室目前配备了双头3D扫描仪、虚拟和增强现实设备、强大的3D图形集群和一套高端图形工作站。具体来说，我们将研究：1）计算技术与宏观层面的特征/模式发现技术适用于大数据集的紧密耦合可视化;这形成了本研究的核心，2）将原始3D扫描数据处理成在诸如游戏、电影、工程、安全等应用中容易使用的格式的新方法，3）使用实验室中的图形集群，在CPU和GPU（被视为协处理器）上为非常大的数据集提供功能和数据分布。和4）开发结合这些技术的交互式和沉浸式系统，用于网络使用、生物信息学、患者教育和临床数据的多学科应用。研究结果将极大地有利于图形研究界以及加拿大的游戏，电影，医疗保健和工程行业。