乳腺X线图像肿块分类方法研究
发布时间：2019-05-30 06:57
【摘要】：乳腺X线图像肿块的良恶性分类是计算机辅助诊断的研究内容之一,本文对乳腺X线图像肿块边缘分割及不同特征的肿块良恶性分类进行研究.基于最大化分割后图像类间方差的思想,提出了一种改进控制标记分水岭方法完成粗分割,然后采用无边缘活动轮廓(CV)模型对粗分割结果进行修正.为了验证不同特征在肿块良恶性分类中的性能,对现有形状特征、纹理特征在不同分类器下的分类性能进行测试.实验在开源数据库DDSM上验证,结果表明,在通过自动分割方法得到肿块边缘的情况下,纹理特征的分类性能更好.
[Abstract]:The classification of benign and malignant breast X-ray masses is one of the research contents of computer aided diagnosis. In this paper, the edge segmentation of breast X-ray masses and the classification of benign and malignant masses with different characteristics were studied. Based on the idea of maximizing the inter-class variance of images after segmentation, an improved control mark watershed method is proposed to complete rough segmentation, and then the rough segmentation results are modified by (CV) model without edge active contours. In order to verify the performance of different features in the classification of benign and malignant masses, the classification performance of existing shape features and texture features under different classifiers was tested. The experiment is verified on the open source database DDSM. The results show that the classification performance of texture features is better when the edge of the mass is obtained by automatic segmentation.
【作者单位】： 北京交通大学电子信息工程学院;
【基金】：国家自然科学基金项目(61502025,61571036) 中国博士后科学基金项目(2015M570029) 北京交通大学人才基金(2015RC024) 中央高校基本科研业务费专项资金(2016JBM010)~~
【分类号】：R730.44;R737.9;TP391.41
本文编号：2488621
[Abstract]:The classification of benign and malignant breast X-ray masses is one of the research contents of computer aided diagnosis. In this paper, the edge segmentation of breast X-ray masses and the classification of benign and malignant masses with different characteristics were studied. Based on the idea of maximizing the inter-class variance of images after segmentation, an improved control mark watershed method is proposed to complete rough segmentation, and then the rough segmentation results are modified by (CV) model without edge active contours. In order to verify the performance of different features in the classification of benign and malignant masses, the classification performance of existing shape features and texture features under different classifiers was tested. The experiment is verified on the open source database DDSM. The results show that the classification performance of texture features is better when the edge of the mass is obtained by automatic segmentation.
【作者单位】： 北京交通大学电子信息工程学院;
【基金】：国家自然科学基金项目(61502025,61571036) 中国博士后科学基金项目(2015M570029) 北京交通大学人才基金(2015RC024) 中央高校基本科研业务费专项资金(2016JBM010)~~
【分类号】：R730.44;R737.9;TP391.41
【相似文献】
相关硕士学位论文 前2条
1 顾晓雯;单、双指数扩散加权成像在肺结节/肿块良恶性鉴别中的初步探讨[D];南通大学;2016年
2 侯慧;基于分水岭算法的白细胞分割研究[D];湘潭大学;2016年
，本文编号：2488621
本文链接：https://www.wllwen.com/yixuelunwen/fangshe/2488621.html
