SVM及其在尿有形成分识别中的应用

对尿液显微图像中一些有形成分进行有效分类识别,具有重大临床意义.通过支持向量机(SVM)这种在训练样本数很少的情况下,能达到很好分类推广能力的学习算法,运用统计学习理论和支持向量机相关概念,将支持向量机引入尿显微图像有形成分识别中,采用数字图像处理技术对尿液有形成分显微图像进行目标特征提取.使用SVM分类的实验结果表明,在样本数不多情况下可获得很好分类效果.

SVM and its application in urine sediment recognition

Urine sediment micro image includes tangible composition that has significant clinical value.Classification and recognition of these composition effectively is very important for diagnosis.The support vector machine is a learning algorithm,which has a good classification ability to deal with limited training samples.Using digital image processing to extract urine sediment features,a novel SVM for the urine sediment recognition is proposed in this paper.It is shown that tangible composition in urine can be distilled and classified so that the classification of urine sediment is well improved with this algorithm.