Method for Character Domain Extraction from Image Using Wavelet Transform

被引：0

作者：

Taniguchi, Taiki ^{[1
]}

Yoshitomi, Yasunari ^{[2
]}

机构：

[1] ZENSHO HOLDINGS Co Ltd, Minato Ku, 2-18-1 Konan, Tokyo 1080075, Japan

[2] Kyoto Prefectural Univ, Grad Sch Life & Environm Sci, Sakyo Ku, 1-5 Nakaragi Cho, Kyoto 6068522, Japan

来源：

PROCEEDINGS OF INTERNATIONAL CONFERENCE ON ARTIFICIAL LIFE AND ROBOTICS (ICAROB2015) | 2015年

关键词：

Internet; Private information; URL of illegal Web; Character domain extraction; Wavelet Transform; Image processing;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The number of images having private information and/or URL of illegal Web site has been increasing in the cyber space on the Internet. These images might cause infringement of human right and/or crime. In the present study, a method for extracting the region(s) having characters on an image has been developed using the discrete wavelet transform and the empirical knowledge that a character has strong vertical and/or horizontal element(s). The experimental results show the usefulness of the proposed method.

引用

页码：302 / 305

页数：4