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

Research and Reviews : Journal of Computational Biology

Development of Intensity-based Segmentation Technique for Meningioma Tumor Detection in MRI Images

  1. Vanshika Gupta
  2. Shaurya Tomer
  3. Nikita Jain

Abstract

There are many types of brain tumors. Some brain tumors detection system using segmentation and classification of MRI images. Brain tumors can have any shape or cut. This encourages us to use high-capacity deep neural networks. Segmentation task and 8000 images for classification task of our neural network and found the best architecture to use. convolutional neural network. In recent years, the three most common forms of brain tumours—glioma, meningioma, and pituitary—have been detected and classified most frequently using deep transfer learning (TL) techniques. The inceptionresnetv2 TL algorithm, which is the best classification algorithm, performs better and gets a superior level of accuracy in identifying and categorizing glioma, meningioma, and pituitary brain cancers, based on the findings of the fine-grained classification experiment. We compare hybrid approaches—which employ an SVM for classification and a CNN for deep feature extraction—with the effectiveness of the inceptionresnetv2 TL algorithm to validate and guarantee the performance of TL classifiers.

Keywords

Support