基于PCA-SVM模型的红斑鳞屑性皮肤病识别研究  

Research of Erythemato-squamous Diseases Recognition Based on PCA-SVM Model

作　　者：徐明月 林泽轩 顾彦 XU Mingyue;LIN Zexuan;GU Yan(School of Computer,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China)

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2018年第6期35-40,共6页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：为提高多特征信息重叠情况下的红斑鳞屑性皮肤病识别准确率,提出一种基于主成分分析法与支持向量机相结合的识别分类方法。结合临床学与病理学选取34个特征指标,对其进行主成分分析法降维处理,降维处理后形成12个综合指标,将处理后训练样本作为支持向量机的输入并建立分类识别模型。仿真实验结果表明,该方法对银屑病、脂溢性皮炎、扁平苔藓、玫瑰糠疹、慢性皮炎和毛发红糠疹等6种红斑鳞屑性皮肤病的识别准确率最高达96.0%,较支持向量机方法和反向传播算法方法,平均准确率分别提升13.9%和17.2%。To improve the recognition accuracy of erythemato-squamous diseases under multi-feature information overlapping,a recognition and classification method based on PCA and SVM is proposed.According to the clinical medicine and pathology,34 features are selected for the diagnosis.After the PCA dimensional reduction,12 comprehensive indexes are formed,which are used as the inputs of SVM to establish the classification recognition models.The results of the simulation show that the accuracy of PCA-SVM method is up to 96.0% for the identification of six types of erythematosquamous diseases with psoriasis,seborrheic dermatitis,lichen planus,pityriasis rosea,chronic dermatitis and pityriasis rubra pilaris,and the average accuracy was 13.9% and 17.2% higher than that of SVM and BP respectively.

关 键 词：多特征 红斑鳞屑性皮肤病 主成分分析法 支持向量机 降维 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]