TY  - JOUR
T1  - MRI Technique Based Detection and Classification of Brain Tumor using
Support Vector Machine (SVM) and k-Nearest Neighbor (kNN)
AU - Jassim Motlak, Hassan 
JO  - Journal of Engineering and Applied Sciences
VL  - 14
IS  - 11
SP  - 3625
EP  - 3629
PY  - 2019
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2019.3625.3629
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2019.3625.3629
KW  - MRI technique
KW  -brain tumor
KW  -support vector machine
KW  -k-nearest neighbor
KW  -accuracy
KW  -MATLAB
AB  - This study presents a system has the ability to detect and classify brain cancer effectively and
efficiently based-on processing images that are combined with a Magnetic Resonance Imaging (MRI)
technique. MRI has high features in dealing with human life such as safety, reliability and it is ability to image
in any plane. The proposed system starts with the preprocessing of images includes resizing and enhancement
of gray images of brain tumor. Textures features of the brain tumor are extracted using two algorithms called
GCLM and k-means. The final stage to classify the tumor if benign or malign was accomplished using two
techniques are k-Nearest Neighbor algorithm (kNN) and Support Vector Machine (SVM). The simulation results
using MATLAB environment, showed that the accuracy of SVM classifier was better than kNN in the
classification of brain tumor where the results are 79 and 73%, respectively. But the value of specificity of the
system for kNN method was higher than SVM and the results are 87.5 and 61%, respectively.
ER  -