计算机视觉核心技术现状与展望
许志杰;王晶;刘颖;范九伦
【期刊名称】《西安邮电学院学报》 【年(卷),期】2012(017)006
【摘要】数字图像和视频数据蕴含了丰富的视觉资源,如何智能化地提取和分析其中的有用信息逐渐成为近年的研究热点。计算机视觉（CV）技术在此学术背景下逐渐发展,并已经广泛应用于生产制造、智能安检、图像检索、医疗影像分析、人机交互等领域。与此同时,计算机视觉技术仍然面临诸如语义信息描述模糊、图像特征检测不稳定且效率低下等诸多问题。本文围绕计算机视觉的核心技术讨论了产生这些问题的根源,并基于对最新技术的讨论,描述了其基本理论框架,最后基于上述内容回顾各个时期重要的理论热点并讨论了计算机视觉的发展趋势。%Digital images and video data provide rich information on their content.The effective and efficient extraction and analysis of that information has been the core of computer vision(CV) research.In the last two decades,CV applications have been widely used in automated manufacturing,intelligent surveillance system,image retrieval and management,medical
image
analysis
and
human
computer
interactions.Although substantial progresses have been made in many areas,current CV techniques still suffer from problems such as low accuracy,slow speed and more prominently,semantic ambiguity.This paper provides a brief review of the CV development,core theoretical foundation,key techniques,successful applications as well as main
计算机视觉核心技术现状与展望
