基于色彩聚类的自然场景文本检测

来源期刊:中南大学学报(自然科学版)2015年第6期

论文作者:李香花 吴慧 赵于前 邹北骥 邹润民

文章页码:2098 - 2104

关键词:场景文本检测;色彩聚类;文本验证

Key words:scene text detection; color clustering; text verification

摘    要:基于场景图像背景复杂,图像中的文本变化多样,提出一种基于色彩聚类的自然场景文本检测算法。其步骤为:首先,用K-均值聚类算法将彩色图像聚成5种色彩层,分析各色彩层中连通域外接矩形的几何特性,去掉离散的非文本连通域;然后,根据相邻连通域外接矩形的几何性质和空间位置关系,将离散的连通域连接形成候选文本块;最后,通过分析候选文本块的几何特性和边缘密度验证文本,得到最终文本检测结果。实验结果证明了本文算法的有效性和可行性。

Abstract: Considering that the natural scene images have complex background and the text in them is usually various, a text detection method based on color clustering was proposed. The procedures are as follows. Firstly, K-means was utilized to group pixels in color images into five different color layers and the bounding-boxes of connected-components in every color layer were obtained. Some non-text components were removed by analyzing geometrical properties of the bounding-boxes. Subsequently, the scattered components were connected to form candidate text blocks according to their bounding-boxes’ geometrical properties and locations. Finally, text verification was performed by geometrical and edge density identifications. The experimental results show that the proposed method is feasible and effective.

相关论文

  • 暂无!

相关知识点

  • 暂无!

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号