Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (36): 190-192.

• 图形、图像、模式识别 • Previous Articles     Next Articles

Recognition of extracting Uyghur texts from videos images

ELI Jume1,Halidan.A1,HUANG Hao2   

  1. 1.Department of Electrical Engineering,Xinjiang University,Urumqi 830047,China
    2.Department of Information Science and Engineering,Xinjiang University,Urumqi 830046,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-12-21 Published:2011-12-21

视频图像中维吾尔文字的识别研究

艾力·居麦1,哈力旦·A1,黄 浩2   

  1. 1.新疆大学 电气工程学院,乌鲁木齐 830047
    2.新疆大学 信息科学与工程学院,乌鲁木齐 830046

Abstract: This paper presents a recognition system for automatically extracting text in video images with complex backgrounds,using self-adaptive iterative algorithm extracts Uyghur characters from video image.After preprocessing text images,the system needn’t segment texts into characters,instead,the system extracts statistical features from a window which is sliding text image.Then the resulting feature vectors are input to the Hidden Markov Models(HMM) for training and recognition.

Key words: video image, Uyghur text, sliding window, Hidden Markov Models(HMM)

摘要: 提出了从复杂背景视频图像中提取文字并识别的一套算法,利用自适应迭代算法提取视频中维吾尔文字,针对维吾尔文字的一些特点,利用合适的预处理方法保留维吾尔文字中的各种点及特殊笔画,同时有效地消除了复杂背景带来的噪声。考虑维吾尔文字书写的特点,利用滑动窗口法提取文字特征避免了文字分割,将产生的特征向量输入到隐马尔可夫模型(Hidden Morkov Model)中进行训练和识别。

关键词: 视频图像, 维吾尔文字, 滑动窗口, 隐马尔可夫模型