摘要
针对电子病历中图像分割问题,提出了基于约束关系的改进核聚类算法,该算法通过引入约束关系在图像分割前进行修正,从而提高图像分割效果。该核聚类算法在MR I中电子病历图像分割实验的结果表明,施加约束关系的核聚类算法能有效地解决电子病历图像中含噪声以及灰度不均匀等问题,具有一定的鲁棒性和较好的图像分割效果。
An improved kernel clustering algorithm based on constraints was proposed, which was used to resolve the problem of the image segmentation in the electronic patient record. The proposed algorithm was designed to retouch the images before they were segmented to improve the image segmentation. The experiment adopted the kernel clustering algorithm in MRI image segmentations, and the results show that the proposed algorithm, compared with classical algorithm, is more effective in solving the inherent problems in the electronic patient record images including noises and intensity inhomogeneities, etc. And accordingly the proposed algorithms are robust and can achieve better image segmentation results.
出处
《计算机应用》
CSCD
北大核心
2007年第8期2066-2068,2071,共4页
journal of Computer Applications
基金
江苏省自然科学基金资助项目(02KJB520004)
南通大学自然科学基金资助项目(05Z061)
关键词
核聚类
电子病历
图像分割
鲁棒性
kernel clustering
electronic patient record
image segmentation
robustness
