Author(s):
1. Soobia Saeed:
Department of Software Engineering, Universiti Teknologi Malaysia (UTM), Malaysia
2. Asadullah Shaikh:
Department of Information Systems, Najran University, Saudi Arabia
3. Muhammad A. Memon:
Institute of Information & Communication Technology University of Sindh, Pakistan
4. Syed M. R. Naqvi:
Department of Computer Science, Sheridan College, Canada
Abstract:
Medicinal images assume a key part in the diagnosis of tumors. In a similar way, MRI could be the cutting edge regenerative imaging technology which permits corner sectional perspective of the body, facilitating the specialists to inspect the affected person. The detailed MRI images enable the medical specialists to recognize minor improvements of structures throughout the body, and serve as a basic part in finding and treatment planning. In this paper, we address the issue of classifying a brain MRI image to predict whether it contains a tumorous part or not. We utilize Histogram of Oriented Gradients (HoG) features to train a Support Vector Machine (SVM) to efficiently predict an MRI image to contain a tumorous part. Before applying the trained SVM classifier on the test MRI images, we also perform image enhancement to increase the accuracy of the prediction. Our experimental results show an impressive accuracy of the proposed technique.
Page(s):
36-40
DOI:
DOI not available
Published:
Journal: Quaid-e-Awam University Research Journal of Engineering, Science and Technology, Volume: 16, Issue: 1, Year: 2018
Keywords:
image segmentation
,
feature extraction
,
Brain Tumor
,
HoG Histogram of Oriented Gradients
,
Magnetic Resonance Image MRI
,
Image Enhancement
,
SVM support vector machine
References:
References are not available for this document.
Citations
Citations are not available for this document.