Application of Modified K Means Clustering Algorithm in Segmentation of Medical Images of Brain Tumor

The process of segmenting medical images serves as a vital technique in partitioning the image into different clusters or homogeneous regions. Lots of techniques and algorithms were developed and applied in various applications. Magnetic Resonance Images (MRI) are used for producing images in the so...

Full description

Saved in:
Bibliographic Details
Published in:Biosciences, biotechnology research Asia Vol. 14; no. 2; pp. 735 - 739
Main Authors: Iqbal, Sana, Sheetlani, Jitendra
Format: Journal Article
Language:English
Published: Bhopal Biosciences Biotechnology Research Asia 28.06.2017
Subjects:
Algorithms
Artificial intelligence
Biopsy
Brain
Brain cancer
Brain research
Brain tumors
Cell division
Cluster analysis
Clustering
Computational neuroscience
Data mining
Error analysis
Hydrogen
Image processing
Image segmentation
Magnetic fields
Magnetic resonance imaging
Medical imaging
NMR
Noise reduction
Nuclear magnetic resonance
Signal to noise ratio
Soft tissues
Software
Vector quantization
ISSN:0973-1245, 2456-2602
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The process of segmenting medical images serves as a vital technique in partitioning the image into different clusters or homogeneous regions. Lots of techniques and algorithms were developed and applied in various applications. Magnetic Resonance Images (MRI) are used for producing images in the soft tissues of human body. The presence of noise in the MRI images of Brain is a multiplicative factor and the reduction of noise is required to obtain good quality in segmentation. However, the concept of accurate segmentation in MRI images is more important and crucial for the proper diagnosis by computational tools aided to perform clinical studies. More clustering algorithms were developed for the segmentation of images from magnetic resonance. However most of them have their limitations and in order to overcome those limitations, a modified version of k means clustering methodology is proposed. The comparison of existing approaches in segmentation such as C-Means Clustering and K-Means Clustering with the Modified version of K Means Clustering is performed to evaluate the performance. Finally certain outcomes were generated in the clustering algorithm of Fuzzy c- means, k-means and modified version of k means for MRI taken in brain and it was observed that the modified version of clustering technique in Kmeans gives better results for the complete performance by measuring parameters such as the index measure of structural similarity, content of structure, mean squared error and analysis of signal noise ratio.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0973-1245
2456-2602
DOI:10.13005/bbra/2502