Detection Of Brain Tumor Using Intelligent Image Classification For Medical Datasets

Abstract

Brain tumor is the most adverse diseases that influence maximum individuals consisting of
youngsters too. The possibility of existence could increase if the tumor is detected at its earlier
phase. In this study, a novel approach is suggested named as MRI aided brain tumor identification
and feature extraction by means of FTS and SVM. Initially, in this approach, the local texture data of
a pixel is achieved through fuzzy texture unit (FTU) and universal texture data is achieved through
fuzzy texture spectrum (FTS). The intention of this study is to present the utility of FTS for texture
Segmentation. Moreover, to enhance the accurateness and quality rate of the SVM aided categorizer,
appropriate characteristics are considered from every segmented tissue. The obtained outcomes of
suggested approach are verified and authenticated for efficiency and quality examination of MRI
brain images, depending on ssensitivity, accuracy, specificity, and dice similarity index. The
experimental outcomes attained97.72% of sensitivity, 96.51% of accuracy and 94.2% of specificity,
and determining the firmness of suggested methodology for detecting the usual and unusual tissues of
brain MR images. The experimental outcomes demonstrated that the strength of the suggested
approach in addition to the enhancements is above the prevailing supervised feature selection
approach.