Evaluation of Segmentation in Magnetic Resonance Images Using k-Means and Fuzzy c-Means Clustering Algorithms
The purpose of cluster analysis is to partition a data set into a number of disjoint groups or clusters. Members within a cluster are more similar to each other than to members from different clusters. Applicability of the centroid-based ¿-means and representative object-based fuzzy c-means algorith...
Saved in:
Published in | Elektrotehniski Vestnik Vol. 79; no. 3; p. 129 |
---|---|
Main Author | |
Format | Journal Article |
Language | English Slovenian |
Published |
Ljubljana
Elektrotehniski Vestnik
01.01.2012
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The purpose of cluster analysis is to partition a data set into a number of disjoint groups or clusters. Members within a cluster are more similar to each other than to members from different clusters. Applicability of the centroid-based ¿-means and representative object-based fuzzy c-means algorithms for study of the Magnetic Resonance Images is analysed in the work. The two algorithms are implemented and their applicability for the analysis of the MRI is evaluated. The criterion is the quality of the clustering compared to the clusters in the reference images. The quality of clustering in both algorithms depends on the number of data points in the image and on the number of the to-be-formed clusters. The algorithms are implemented, they are applied to the reference two-dimensional multispectral MR images and the resulting image segmentation is analysed by the objective criteria. The object-based fuzzy c-means algorithm outperforms the centroid-based ¿-means algorithm by all means. The advantage of the former stems from utilization of the cluster membership fuzziness. [PUBLICATION ABSTRACT] |
---|---|
Bibliography: | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
ISSN: | 0013-5852 2232-3236 |