欢迎访问《图学学报》 分享到:

图学学报

• 视觉与图像 • 上一篇    下一篇

基于改进的半监督FCM 聚类算法的肺结节分类与识别

  

  • 出版日期:2015-04-30 发布日期:2015-06-03

Modified Fuzzy Clustering with Partial Supervision Algorithm in Classification and Recognition of Pulmonary Nodules

  • Online:2015-04-30 Published:2015-06-03

摘要: 对肺结节的分类识别是肺部肿瘤计算机辅助诊断系统的关键环节。为了提高肺结
节分类识别的准确率,针对肺结节的病变特征提取出一组以形状特征为主的特征向量,同时基
于LIDC 数据库中医生提供的标记信息,提出一种改进的半监督FCM 聚类分析算法,利用部分
标记样本的类别信息来指导聚类过程,使非标记样本更准确的聚类。实验结果表明,本文方法
能得到更高的分类准确率。

关键词: 计算机辅助诊断, 半监督FCM 聚类, 病变特征, 标记信息

Abstract: Accurate classification and recognition of pulmonary nodules is an important and key
process of lung cancer computer-aided diagnosis system. In this paper, to improve the accuracy, we
propose a modified partial supervised fuzzy clustering algorithm based on the annotation information
of doctors in LIDC database. First, all pulmonary nodules are segmented from the CT images. Second,
according to the lesion characteristics of pulmonary nodules, we extract a set of mainly shape-based
feature vectors. Finally, we calculate the reference membership by exploiting the class information of
labeled samples in the process of clustering, and use the reference membership to guide the clustering
process of the testing samples, for helping the testing samples to cluster more accurate. Experimental
results show that the proposed method can out-perform the traditional algorithm in classification and
recognition.

Key words: computer-aided diagnosis, fuzzy C-means clustering with partial supervision, lesion
characteristic,
annotation information