基于LLE和LS_SVM的胃粘膜肿瘤细胞图像分类

胃粘膜肿瘤细胞图像的复杂性,组织器官形状的不规则性以及不同细胞的差异性,使得采用一般的线性分类方法对其进行分类很困难,结合局部线性嵌入（LLE）在处理非线性数据及最小二乘支持向量机（LS_SVM）在处理小样本、高维数及泛化问题方面的优势,文章提出一种基于LLE＋LS_SVM的胃粘膜肿瘤细胞图像分类方法,并采用LS_SVM的线性拟合误差来判断实验效果,最后比较本文方法与其他分类方法的优越性。实验结果表明,该方法在分类准确率和运行时间方面都有很大的优势。

Classification of Gastric Cancer Cells Based on LLE and LS SVM

It is difficult to recognize gastric tumor cell images by the the linear classification methods for the complexity of gastric tumor cell images,the irregular shape of tissues and organs and the differentiation of different cells.As nonlinear classification methods,local linear embedding（LLE） can well deal with nonlinear data and least squares support vector machine（LS_SVM） can well resolve small sample size,high dimension and generalization issues.A classification method is proposed in this paper based on LLE and LS_SVM.The linear fitting function is used to fit its linear errors,the linear fitting error is used to determine the results,finally superiority of method in this paper is compared with other classification methods.It is proved by the experiment results that this method has a significant advantage in classification accuracy and running time.