【作者】 王伟强； 付立波； 高文； 黄庆明； 蒋树强；

【Author】 Weiqiang Wang~(1,2) Libo Fu~1 Wen Gao~1 Qingming Huang~(1,2) Shuqiang Jiang~1 (1.Affiliation 1,City 1 Post Code 1)、 (2.Affiliation 2,City 2 Post Code 2)

【机构】 中国科学院计算技术研究所；

【摘要】 图像中的叠加文字是理解图像语义的重要语义线索。叠加文字区域除对比度强和边缘密集特点外,文字笔画的线条、尺度以及空间分布等也会随着不同语言而呈现出不同于背景区域的特定纹理模式。本论文提出一种利用笔画线条的统计特征基于支持向量机进行图像中叠加文字检测的方法。以汉字为例的初步实验表明所提出笔画纹理特征对较多字符构成的文字区域具有很好的性能。更多还原

【Abstract】 Embedded text in images is an important clue for understanding image semantic information.Besides strong contrast and dense edges,embedded text has a specific texture pattern different from background,due to line,scale and spatial distribution of character strokes from different languages。This paper presents a text detection method based on support vector machine(SVM)using the statistics features characterizing character strokes.Our experiments on Chinese characters demonstrate the proposed stroke texture features have good distinction power,especially for text regions composed of many characters。更多还原