基于K-P算法优化的手写汉字细化算法  

Handwritten Chinese character thinning algorithm based on K-P algorithm optimization

作　　者：范勇峰 李成城 林民 FAN Yong-feng;LI Cheng-cheng;LIN Min(College of Computer Science and Technology,Inner Mongolia Normal University,Hohhot 010022,China)

机构地区：[1]内蒙古师范大学计算机科学技术学院,内蒙古呼和浩特010022

出　　处：《计算机工程与设计》2023年第10期3076-3083,共8页Computer Engineering and Design

基　　金：内蒙古自治区科技计划基金项目(2020GG0167);内蒙古自然科学基金项目(2021LHMS06010);242信息安全计划基金项目(2019A114)。

摘　　要：为解决现有汉字细化算法中的骨架断裂、交叉点畸变等问题,提出一种K-P(Kmeans++-PCA)算法,优化现有的手写汉字细化算法。提出交叉点匹配模板,减少骨架毛刺去除时对非交叉点的遍历,借助主成分分析算法与端点距离对骨架断点进行连接判断,对骨架交叉点畸变进行处理,使用点到边界方向距离算法提取交叉区域后结合Kmeans++聚类对畸变区域进行分析,利用笔画走势信息对骨架进行修复并使用B样插值算法完成骨架重建。将优化后的算法在手写汉字数据集上进行实验,验证了优化后的算法优于优化前的算法。To solve the problems of skeleton fracture and intersection distortion in the existing Chinese character thinning algorithms,the K-P(Kmeans++-PCA)algorithm was proposed to optimize the existing handwritten Chinese character thinning algorithms.An intersection matching template was proposed to reduce the traversal of non-intersection points during skeleton burr removal,and the connection judgment of skeleton breakpoints was carried out with the help of principal component analysis algorithm and endpoint distance.The distortion of skeleton intersection points was processed.After the intersection areas were extracted using point to boundary orientation distance algorithm,the distortion areas were analyzed in combination with Kmeans++clustering.The skeleton was repaired using the stroke trend information,and the skeleton reconstruction was completed using B-sample interpolation algorithm.The optimized algorithm was tested on handwritten Chinese character data set,and it is verified that the optimized algorithm is better than the original one.

关 键 词：手写汉字 汉字细化 主成分分析算法 骨架断裂 点到边界方向距离算法 交叉点畸变 Kmeans++聚类 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]