作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2006, Vol. 32 ›› Issue (9): 209-210.

• 多媒体技术及应用 • 上一篇    下一篇

基于多帧分析的视频文本分割和增强

许剑峰,黎绍发   

  1. 华南理工大学计算机学院,广州 510640
  • 出版日期:2006-05-05 发布日期:2006-05-05

Video Text Segmentation and Enhancement Using Multiframe Analysis

XU Jianfeng, LI Shaofa   

  1. College of Computer Science and Engineering, South China Univ. of Tech., Guangzhou 510640
  • Online:2006-05-05 Published:2006-05-05

摘要: 提出了一种将垂直颜色边缘与分块多帧分析的文本定位与增强的方法。根据文本的垂直颜色边缘特征进行粗筛选,获得和跟踪文本候选区;然后通过分块多帧分析对文本进行增强。实验表明,该方法能有效去除复杂背景,提高视频文本的识别率。

关键词: 文本定位;文本增强;分块多帧分析

Abstract: A new method of text location and enhancement based on vertical color edge and block multi-frame analysis is presented. Vertical color edge characteristics of characters are used for locating and tracking candidate text regions first. Then block multi-frame analysis is used to enhance the complex background video text. The experimental results demonstrate that it is efficient in eliminating complex background and increasing the recognition rate of video text.

Key words: Text location; Text enhancement; Block multi-frame analysis