| 聚类核值相似区特征点的医学影像分类 |
| |
| 引用本文： | 李博,曹鹏,栗伟,赵大哲.聚类核值相似区特征点的医学影像分类[J].中国图象图形学报,2013,18(10):1322-1328. |
| |
| 作者姓名： | 李博 曹鹏 栗伟 赵大哲 |
| |
| 作者单位： | 东北大学,东北大学,东北大学,东北大学 |
| |
| 基金项目： | 国家自然科学基金项目（面上项目，重点项目，重大项目） |
| |
| 摘 要： | 针对传统方法通常选取角点或极值点作为特征点，忽略了局部纹理变化从而影响医学影像分类性能的问题，提出一种新的特征点检测和描述方法，并基于其应用Bag-of-Keypoints模型实现医学影像分类。首先改进自适应的K-means对影像进行像素级聚类，构建核值相似区并选取邻域内聚类分布变化急剧的像素点作为特征点，然后在极坐标系中定义特征点描述符并生成视觉词典，通过视觉词直方图描述影像，最后利用直方图交集方法度量影像间的相似度来扩展KNN完成分类。遵循IRMA的医学影像类别编码标准严格选择实验数据，结果表明该算法较传统方法F1值平均提高4.5%，对于不同类别影像效果更加稳定鲁棒，从而更好地满足临床应用需求。
|
| 关 键 词： | 特征点 Bag-of-Keypoints模型 自适应聚类 核值相似区 |
| 收稿时间： | 2012/10/29 0:00:00 |
| 修稿时间： | 2013/1/29 0:00:00 |
Medical image classification based on cluster univalue segment assimilating nucleus feature points |
| |
| Li Bo,Cao Peng,Li Wei and Zhao Dazhe.Medical image classification based on cluster univalue segment assimilating nucleus feature points[J].Journal of Image and Graphics,2013,18(10):1322-1328. |
| |
| Authors: | Li Bo Cao Peng Li Wei and Zhao Dazhe |
| |
| Affiliation: | College of Information Science and Engineering, Northeastern University, Shenyang 110004, China;Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110004, China;College of Information Science and Engineering, Northeastern University, Shenyang 110004, China;Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110004, China;College of Information Science and Engineering, Northeastern University, Shenyang 110004, China;Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110004, China;College of Information Science and Engineering, Northeastern University, Shenyang 110004, China;Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110004, China |
| |
| Abstract: | |
| |
| Keywords: | image classification feature point Bag-of-Keypoints model adaptive cluster USAN |
|
| 点击此处可从《中国图象图形学报》浏览原始摘要信息 |
|
点击此处可从《中国图象图形学报》下载免费的PDF全文 |
|
