首页|基于快速鲁棒模糊C有序均值聚类的苗族服饰图像分割算法

基于快速鲁棒模糊C有序均值聚类的苗族服饰图像分割算法

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
苗族服饰图像具有绣线纹理复杂、色彩形状多样等特征,针对模糊 C有序均值(Fuzzy C-Ordered-Means,FCOM)聚类算法在进行苗族服饰图像分割时,存在耗时长、分割效果不理想的问题,提出了一种快速鲁棒模糊C有序均值聚类算法.在FCOM算法基础上加入了竞争学习的思想,通过构造新的隶属度约束函数,对像素点进行更加强制清晰的划分,提高图像像素定位的准确性,从而加快算法的收敛速度.结果表明,本文算法在图像分割过程中能有效地降低异常值的影响,获得更加准确的分割结果.该算法在Jaccard相似系数、分割精度、Dice相似系数、模糊划分系数及模糊划分熵等性能方面均优于其他几种模糊C均值(Fuzzy C-Means,FCM)算法,且分割时间与迭代次数也优于FCOM算法.
Miao costume image segmentation algorithm based on fast robust fuzzy C-ordered-means clustering
Miao costume images are characterized by complex embroidery texture,diverse colors and shapes.A fast robust fuzzy C-ordered-means clustering algorithm was proposed to solve the problems of long time and unsatisfactory segmentation effect in Miao clothing image segmentation for Fuzzy C-Ordered-Means(FCOM)clustering algorithm.Based on the FCOM algorithm,the idea of competitive learning was added,by constructing a new membership constraint function,the pixels were divided more forcibly and clearly,the accuracy of image pixel positioning was improved,and the convergence speed of the algorithm was accelerated.The results show that the proposed algorithm can effectively reduce the influence of outliers in image segmentation and obtain more accurate segmentation results.The algorithm is superior to other Fuzzy C-Means(FCM)algorithms in Jaccard similarity coefficient,segmentation accuracy,Dice similarity coefficient,fuzzy partition coefficient and fuzzy partition entropy,and the segmentation time and iteration times are also superior to FCOM algorithm.

Miao image segmentationclustering algorithmFuzzy C-Ordered-Meanscompetitive learningrobustness

陈阳、黄成泉、雷欢、彭家磊、覃小素、周丽华

展开 >

贵州民族大学 数据科学与信息工程学院,贵州 贵阳 550025

贵州民族大学工程技术人才实践训练中心,贵州 贵阳 550025

苗族图像分割 聚类算法 模糊C有序均值 竞争学习 鲁棒性

国家自然科学基金贵州省省级科技计划贵州省研究生教育教学改革重点项目贵州省教育厅自然科学研究项目

62062024黔科合基础-ZK[2021]一般342黔教合YJSJGKT[2021]018黔教技[2022]015

2024

10.19333/j.mfkj.20230603809

毛纺科技
中国纺织信息中心 北京毛纺织科学研究所

毛纺科技

北大核心
影响因子：0.3
ISSN：1003-1456
年,卷(期)：2024.52(8)