摘要
目前对高噪声的MRI医学图像还没有较好的分割算法。在研究加强模糊C均值聚类(EnFCM)算法的基础上,提出一种改进算法。首先通过字典学习算法进行降噪,然后利用聚类有效性确定最佳聚类数目,最后利用改进EnFCM算法完成分割。实验表明,该方法能够实现自动分类功能,达到了有效分割高噪声MRI医学图像的目的。
A better segmentation algorithm is not given about high noise MRI medical image at the moment. This paper proposes an improved algorithm based on enhanced fuzzy C-means clustering algorithm(EnFCM). First, we use dictionary learning algorithm to reduce the original image noise, then we use clustering validity to determine the optimal number of clusters, and finally use EnFCM algorithm to complete segmentation. The experiment shows that this method can realize automatic classification function and achieve the purpose of effectively segmenting high noise MRI medical images.
引文
[1]孙黎明.图像稀疏去噪算法的并行改进研究[D].重庆:重庆大学,2011.
[2]姜鹏飞.基于稀疏表示与字典学习的图像去噪算法研究[D].西安:西安电子科技大学,2011.
[3]苗加庆.自适应字典改进Canny算子CT图像分割[J].电视技术,2015,39(15):24-27.
[4]Mallat S,Zhang Z.Matching pursuit with time-frequency dictionaries[J].IEEE Transactions on Signal Processing,1993,41(12):3397-3415.
[5]高颖,王阿敏,王凤华,等.改进的小波变换算法在图像融合中的应用[J].激光技术,2013,37(5):690-695.
[6]杨悦.基于模糊聚类及水平集方法的图像分割技术研究[D].长春:吉林大学,2011.
[7]Szilagyi L,Benyo Z,Szilagyi SM,et al.MRbrain image segmentation using an enhanced fuzzy C-means algorithm[J].Engineering in Medicine&Biology Society Proceedings of Annual International Conf,2003(1):724-726.
[8]陈刚,陆媛,杨慧中.基于小波去噪和改进的FCM算法的基因表达数据分析[J].计算机与应用化学,2011,28(7):952-953.