STM Journals Peer-reviewed research from Consortium e-Learning Network

Research & Reviews: A Journal of Bioinformatics

Brain Tumor Detection

  1. Mahendrakumar Chaudhary
  2. Muskan Lamba
  3. Sanket Datir
  4. Reshma Gulwani

Abstract

The MRI, or magnetic resonance imaging, is one of the most common practices in detecting a brain tumor since it contains important information used to extensively scan the human brain’s internal anatomy. The treatment of any tumor depends entirely on the knowledge and expertise of the physician. This paper reflects the identification of brain tumors using different machine learning algorithms along with the convolution neural network (CNN) and transfer learning. The Histogram of Oriented Gradient (HoG) is used as a function descriptor in machine learning algorithms such as SVM Linear, SVM RBF, random forest, and logistic regression. To enhance brain tumor detection accuracy, a CNN model is combined with pre-trained transfer learning models VGG-16 and ResNet50. Based on experimental results, the logistic regression achieves 95.3% accuracy, while CNN, VGG-16, and ResNet50 attain 84.31%, 94%, and 93.7% accuracy, respectively, which is a compelling measure.

Keywords

Support