摘要
针对因红细胞重叠以及细胞试剂里含有杂质而造成红细胞识别不准确的问题,提出了一种基于支持向量机(Support Vector Machines,SVM)与改进分水岭算法的红细胞识别算法.算法首先提取原始图像的二值化图像中每个连通区域的形状特征;利用SVM识别单个连通区域与重叠连通区域;利用改进分水岭算法将重叠区域分割为单个连通区域;其次提取单连通区域对应原始图像中的颜色特征,利用另一个SVM识别细胞与杂质;最后统计细胞个数.实验结果表明,该算法提升了红细胞计数的准确率.
Aiming at the inaccuracy of red blood cell recognition caused by red blood cell overlap and impurity in cell reagent,a red blood cell recognition algorithm based on support vector machines(SVM)and improved watershed algorithm is proposed.First,the shape features of each connected region are extracted in the binarization image of the original image.SVM is used to identify single connected regions and overlapping connected regions.Second,the overlapping region is divided into a single connected region by the improved watershed algorithm.Third,the color features of the original image is extracted according to the simply connected region,and SVM is used to identify cells and impurities.Fourth,the algorithm count the number of cells.Experimental results show that the algorithm improves the accuracy of red blood cell count.
作者
刘生富
张鹏程
周广宇
刘祎
桂志国
LIU Shengfu;ZHANG Pengcheng;ZHOU Guangyu;LIU Yi;GUI Zhiguo(Shanxi Key Laboratory of Biomedical Imaging and Imaging Big Data, North China University, Taiyuan 030051, China)
出处
《测试技术学报》
2022年第1期48-53,共6页
Journal of Test and Measurement Technology
基金
国家自然科学基金资助项目(11605160)
山西省自然科学基金资助项目(201901D211246)
山西省回国留学人员科研资助项目(2016-089)。
