计算机与数字工程2024，Vol.52Issue(6)：1848-1853,6.DOI:10.3969/j.issn.1672-9722.2024.06.042
场景图像弯曲文本检测与识别
Detection and Recognition of Curved Text in Scence Image
刘航源1
作者信息
- 1. 中国石油大学(华东)计算机科学与技术学院 青岛 266580
- 折叠
摘要
Abstract
With the development of deep learning technology,scene text detection and recognition has made tremendous prog-ress in recent years,and the implementation methods of scene text detection and recognition have become more and more diverse.However,scene text is still mostly represented by rectangles or rotated rectangles.It is difficult to accurately represent curved text instances in these types of text representation.In order to solve the problem of curved text representation,this paper proposes a para-metric curved text representation method,and designs a curved text correction algorithm based on the parametric representation method,which realizes the morphological correction of curved text and reduces the difficulty of recognition.By introducing this method,a fully functional scene text detection and recognition system is realized.The experimental results show that this method can effectively detect and recognize scene text.关键词
场景文本检测/文本识别/弯曲文本/形态纠正Key words
scene text detection/text recognition/curved text/morphological correction分类
信息技术与安全科学引用本文复制引用
刘航源..场景图像弯曲文本检测与识别[J].计算机与数字工程,2024,52(6):1848-1853,6.
