摘要
针对近些年来,女性人群中乳腺恶性肿瘤患者逐年增长的趋势,提出了一种融合纹理特征和形状特征的乳腺超声肿瘤图像的识别方法。该方法不但结合了临床医生的诊断经验,而且有效利用了数字图像特征提取技术,提取出的特征能反映出良性肿瘤和恶性肿瘤的本质区别,将样本进行特征提取并通过支持向量机(SVM)技术分类,该方法取得了良好的分类效果。实验结果证明文中的特征融合方法十分有效。
In recent years, the patients number of female population who have malignant breast tumor have grown very fast. This paper proposed a classification method for breast tumor ultrasound image which fusion texture features and shape features. This method not only combines the experience of the cliniciang diagnosis, but also effectively makes use of the technology of digital image feature extraction. The extracted features reflect the essential difference of benign and malignant breast tumors. Through the feature extraction of samples and the support vector machine （SVM） classification algorithm, the method obtained a good classification effect. The experimental result shows that the feature fusion method is very effective in this paper.
出处
《信息技术》
2013年第6期44-46,50,共4页
Information Technology
基金
黑龙江省自然科学基金项目(F201234)
黑龙江大学高层次人才(创新团队)支持计划(Hdtd2010-07)
黑龙江大学青年科学基金项目(QL201028)
关键词
支持向量机
医学超声图像
特征提取
support vector machines
medical ultrasound image
feature extraction
