Multi-Modal Case Study on MRI Brain Tumor Detection Using Support Vector Machine, Random Forest, Decision Tree, K-Nearest Neighbor, Temporal Convolution & Transfer Learning

Abstract

In the Medical field, Brain Tumor Detection has
become critical and demanding task because of their several
shapes, locations, and the intensity of image. That’s why an
automated system is important to aid physicians and radiologists
in detecting and classifying brain tumor. In this research, we
have discussed different machine learning as well as deep
learning algorithm which are mostly used for image
classification. We have also compared different models that are
being used for tumor classification based on machine learning
and deep learning. MRI images of Glioma tumor, Pituitary
tumor, Meningioma tumor are the base of this research, and we
have compared different techniques along with the accuracy of
different classification model using those MRI images. We have
used different deep learning pre-trained model for training brain
tumor images. Those pre-trained models have provided
outstanding performance along with less power consumption and
computational time. EfficientNet-B3 has provided the best
accuracy of 98.16% among other models as well as traditional
machine learning algorithms. The experimental result of this
model is proven the best and most efficient for tumor detection
and classification in comparison with other recent studies.