一种面向最近邻分类器和动态Snake模型的血红细胞图像特征提取方法  

FEATURE EXTRACTION METHOD OF RED BLOOD CELL BASED ON ACTIVE SNAKE MODEL AND NEAREST NEIGHBOR CLASSIFIER

作　　者：黄伟[1] 周鸣争[1] 赵森严[1] 

机构地区：[1]安徽工程大学计算机与信息学院,安徽芜湖241000

出　　处：《井冈山大学学报（自然科学版）》2016年第3期41-47,共7页Journal of Jinggangshan University (Natural Science)

基　　金：安徽省自然科学研究基金项目(1608085QF154);安徽工程大学计算机应用技术重点实验室开放基金资助项目(JSJKF201503)

摘　　要：利用动态可变的Snake活动轮廓模型,并提取相应的特征区域色彩信息,结合改进的最近邻(Nearest Neighbor,NN)分类器核鉴别算法,通过信息融合方法实现对血红细胞图像的特征精确提取,这种方法可应用于临床辅助检测和诊断过程,具有一定的应用价值。Recently Snake model has been widely used on image extraction and recognition. The key is feature extraction of the contour. The important problems focus on algorithm efficiency and misextraction. NN Classifier is also used with its simplicity. We describe the Snake Model on Red Blood Cell image. A method is proposed that we use improved NN Classifier and color information to detect the contour of feature extraction. The method can be effectively applied in the process of clinical diagnosis.

关 键 词：SNAKE模型 NN分类器 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]