Classification of Brain Lesion using K- Nearest Neighbor technique and Texture Analysis

Texture is an important property for analyzing many types of images. It provides a rich source of information about the image. In this paper four statistical features (contrast, correlation, homogeneity and energy) are calculated from the gray level co-occurrence matrix in which the gray level of va...

Full description

Saved in:
Bibliographic Details
Published inJournal of physics. Conference series Vol. 1178; no. 1; pp. 12018 - 12026
Main Authors Ali, Alyaa H., Al-hadi, Sara Abd, R. Naeemah, Maysaa, N. Mazher, Alaa
Format Journal Article
LanguageEnglish
Published Bristol IOP Publishing 01.02.2019
Subjects
Brain
brain cancer
Brain stroke
Cancer
co-occurrence matrix
Computed tomography
Hemorrhage
Homogeneity
Image classification
Image contrast
K-NN
Medical imaging
Statistical methods
Texture
Online AccessGet full text

Cover

More Information
Summary:Texture is an important property for analyzing many types of images. It provides a rich source of information about the image. In this paper four statistical features (contrast, correlation, homogeneity and energy) are calculated from the gray level co-occurrence matrix in which the gray level of variation associated with gray of three samples from CT scan Images to patients with (hemorrhage, ischemic and cancer) in the brain, the region of interest has been obtained by classifying the images using the k-nearest neighbor (K-NN) five classes has been proposed for the classification purpose, the result shows two classes have infection region and the three other classes are healthy region. It has found that the cancer texture are highly correlated comparing with the stroke. The stroke texture is more homogenous than the cancer texture this make the distinguish between them more specific easy.
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/1178/1/012018