Page 84 - ISMCON souvenir 2021

ISMSCON - 2021

OS57: Brain tumor classification using Support Vector Machine
Learning Technique

Ms. Sivakami Sundari S 1
1Divison of Epidemiology and Biostatistics, St. John’s Research Institute, Bengaluru.
Mail Id of the presenting author: sivakami.s@sjri.res.in

Abstract:
Background:

At present, processing of medical images is a developing and important field. Tumor detection imaging
techniques are not limited to Computed Tomography scans (CT scans), X-rays and Magnetic Resonance
Imaging (MRI), Positron Emission Tomography (PET) etc. Brain tumor identifications through MRI images
is a difficult task because of the complexity of the brain. MRI images can be processed and the brain tumor
can be segmented. These tumors can be segmented using various image segmentation techniques.
Objectives: To detect meningioma and pituitary tumor on an MRI image using support vector machine
learning and logistic regression technique.
Methods: The MRI image dataset were acquired from an online open-source Kaggle medical image
repository. In this study images were generated using support vector machine (SVM) learning and logistic
regression technique. The SVM algorithm is based on the study of a supervised learning technique and
is applied to one-class classification problem to n-class classification problems. The principle aim of the
SVM algorithm is to transform a nonlinear dividing objective into a linear transformation using a function
called SVM’s kernel function. In this study, I used the 2-dimension kernel function to identify the support
vector classifier. By using a kernel function, the nonlinear samples can be transformed into a high-
dimensional future space where the separation of nonlinear samples or data might become possible,
making the classification convenient. Python 3.10 software was used for this study analysis.
Results: The SVM technique identified and marked brain tumors with an accuracy of 98%. Meningioma
and pituitary images are classified better using SVM learning technique than logistic regression. Logistic
regression overestimates the meningioma and pituitary tumor images.

Conclusion: Support vector machine learning technique can be satisfactorily used for classification for
meningioma and pituitary images. Overall, this process could enhance the detection of small, early stage
or occult brain tumors overlooked by the radiologist.

Key words; MRI, Support vector machine learning technique, classification of images, kernel
function.

OS58: CERVICAL CANCER INCIDENCE AND MORTALITY IN
SOUTH EAST ASIA: EVIDENCE FROM GLOBOCAN 2020

Smita Asthana*, Satyendra Singh Yadav, Shalini Singh
Affiliation: ICMR-National Institute of Cancer Prevention & Research, I-7, Sec 39, Noida, 201301
Email: smitasanjay97@yahoo.com
Abstract
Aim- Cervical cancer is one of the leading malignancies among females in SouthEastAsian region
(SEAR) as defined by WHO. This study aims to examine the cervical cancer burden in SEAR using
recently released Globocan 2020 estimates.

Study Design- Ecological

82 CONFERENCE SOUVENIR