计算机工程与应用2011，Vol.47Issue(7)：189-192,4.DOI:10.3778/j.issn.1002-8331.2011.07.055
改进的视频帧中文本区域定位算法
Improved algorithm for text region localization in video frames.
摘要
Abstract
Contraposing the characters of images in video, such as complicated backdrop and big diversification of font,this paper proposes an improved algorithm for text region localization in video. The algorithm designs and implements a detecting cross points algorithm at first. Then the algorithm recognizes and deletes most of non-text edges with the features of edge like the density of cross points,in order to reduce the influence of backdrop on text region,and then expands the rest edges to form candidate text regions. The text region and non-text region are distinguished by combing the features of text region and a SVM classifier. The experiments show that the algorithm can extract more than 90％ of text regions contained in video frames and the precision can reach up to 92.0％.关键词
边缘检测/交叉点检测/图像膨胀/支持向量机Key words
edge detection/cross point detection/image expand/support vector machine分类
信息技术与安全科学引用本文复制引用
张海云,张凤荔,刘伟,万明成..改进的视频帧中文本区域定位算法[J].计算机工程与应用,2011,47(7):189-192,4.基金项目
国家高技术研究发展计划(863)(the National High-Tech Research and Development Plan of China under Grant No.2006AA01Z41). （863）
