1 Image Processing
1 Image Processing
复制标题
DOI：
--
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
Tom Funkhouser
中科院分区：
文献类型：
--
作者：
Tom Funkhouser
图像处理的目标是从图像和视频中稳健地提取有用的高级信息。有用的高级信息的类型取决于应用程序。应用示例包括用于监视的对象检测和跟踪、用于自动化生产系统的缺陷检测以及用于遥感和地图注释的场景分类。提取高级信息是一个困难的研究问题，并且已经提出了许多不同的算法。自 20 世纪 90 年代初成立以来，图像和视频处理一直是可重构计算社区的重要应用驱动因素。在现代可重构设备出现之前，图像和视频处理也是计算机体系结构研究和超大规模集成电路设计的重要应用驱动因素。该应用领域在计算机体系结构社区中受到如此多的关注有两个相关原因。第一个是在通用处理器实现中观察到的较差的、通常不可接受的性能。这可以归因于：
The goal of image processing is to robustly extract useful, high-level information from images and video. The type of high-level information that is useful depends on the application. Examples of applications include object detection and tracking for surveillance, defect detection for automated production systems, and scene classification for remote sensing and map annotation. Extracting high-level information is a difficult research problem and many different algorithms have been suggested. Image and video processing have been significant application drivers for the reconfigurable computing community since its inception in the early 1990’s. Prior to modern reconfigurable devices, image and video processing were also significant application drivers for computer architecture research and VLSI design. There are two related reasons why this application domain has received so much attention in the computer architecture community. The first is the poor, often unacceptable, performance observed in general-purpose processor implementation. This can be attributed to: