Author : Dr. O.A. Mohamed Jafar 1
Date of Publication :23rd January 2018
Abstract: Brain tumor segmentation is a challenging problem. The medical expert can analyze the MRI (Magnetic Resonance Imaging) but this activity is error-prone and time consuming process. Data clustering is the process of grouping similar objects. Existing brain tumor segmentation methods use Euclidean distance metric. In this paper, a study of hard and fuzzy methods based on different distance metrics – Manhattan and Chebyshev used for brain tumor segmentation is done. The result of Adaptive Fuzzy K-means (AFKM) is compared with K-means and Fuzzy K-means (FKM) algorithms. The AFKM based on Chebyshev distance gives better result than K-means and FKM techniques in terms of accuracy and quantitative measurements.
Reference :