期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2018
卷号：18
期号：1
页码：99-111
出版社：International Journal of Computer Science and Network Security
摘要：In the current era of technology, information acquisition from the images become most important task due to the rapid development of data mining and machine learning. This paper introduces text detection from natural scene images. The paper provides analysis and compares different methods to detect and localize the text in natural scene images. The text exists in an image under varying properties such as text size, font, style, illumination, and orientation with complex background. We have addressed each state-of-the-art methods to resolve these issues. Next, experimental results and performance evaluations of several algorithms are presented. The results carried out under the various text detection approaches in form of precision and recall are given. Furthermore, different datasets are enlisted which are freely and publicly available, and performance protocols are also defined briefly.
关键词：;;;; ;;;;;; Text Detection;Text localization;Methodologies;Survey.
