首页 | 本学科首页   官方微博 | 高级检索  
     

基于小波神经网络的图像文本信息提取技术研究
引用本文:黄同城,丁友东. 基于小波神经网络的图像文本信息提取技术研究[J]. 湖南师范大学自然科学学报, 2006, 29(2): 24-27
作者姓名:黄同城  丁友东
作者单位:1. 上海大学计算机工程与科学学院,中国,上海,200072;邵阳学院信息与电气工程系,中国,邵阳,422000
2. 上海大学计算机工程与科学学院,中国,上海,200072
基金项目:上海市科技局基金资助项目(04SHK1068),湖南省教育厅基金资助项目(02C303)
摘    要:提出了利用小波神经网络提取图像中文本信息的新颖方法.原图像经过离散小波变换分解成4个子频带,文本区域的高频子频带与非文本区域的不同,所以可利用其差异计算出3个特征值作为人工神经网络的输入值,然后用基于BP算法构建的人工神经网络来训练待测的文本区域.文本区域的人工神经网络输出值不同于非文本区域的输出值,因此可利用一阈值来判定其是否为文本区域.最后,将可检测的文本区域经过扩张运算后便可得到正确的文本区域.

关 键 词:文本提取  小波变换  人工神经网络
文章编号:1000-2537(2006)02-0024-04
收稿时间:2006-03-27
修稿时间:2006-03-27

Text Extraction from Image Based on Wavelet Neural Network
HUANG Tong-cheng,DING You-dong. Text Extraction from Image Based on Wavelet Neural Network[J]. Journal of Natural Science of Hunan Normal University, 2006, 29(2): 24-27
Authors:HUANG Tong-cheng  DING You-dong
Affiliation:1. School of Computer Engineering and Science, Shanghai University, Shanghai 200072, China; 2. Department of Information and Electrical Engineering, Shaoyang University, Shaoyang 422000, China
Abstract:A novel text extraction approach based on wavelet neural network from image or video is presented.It successfully extracts features of candidate text regions using discrete wavelet transform.This is because the intensity characteristic of any detail component sub-band is different from that of the others.Utilizing this difference,features of candidate text regions are extracted.A neural network based on back propagation(BP) algorithm is trained according to these features.The final network output of real text regions is different from those non-text regions.Hence an appropriate threshold value with some dilation operators can be applied to obtain the real text regions.
Keywords:discrete wavelet transform  text extraction  neural network
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号