Scholarlink

Accurate diagnoses of brain tumors are required to be treated as early as possible. In this study, the authors have compared four existing convolutional neural network (CNN) models, Inception V3, ResNet50, ResNet101, and VGG19 in the context of classifying brain tumor based on MRI images. The performance of the models is determined using the classification accuracy where VGG19 has the highest accuracy of 97.00 and ResNet50 has the second highest accuracy of 96.37. Inception3 and ResNet101 had the accuracies of 93.80 and 91.74, respectively. When comparing ResNet50 and VGG19, the deepest structures are found to capture more complicated features, and this makes them best in the classification of brain tumors. These results highlight the significance of proper choice of architecture in deep-learning applications to medical image analysis to improve automated diagnostic systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparative Analysis Brain Tumor Classification Using Different Convolutional Neural Networks (CNN) Algorithms

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Morium Akter Munny,
  • Isha Das,
  • Md Abdul Alim Sarkar,
  • Sayed Sayem,
  • Rashni Akter,
  • Koishik Ahmed

摘要

Accurate diagnoses of brain tumors are required to be treated as early as possible. In this study, the authors have compared four existing convolutional neural network (CNN) models, Inception V3, ResNet50, ResNet101, and VGG19 in the context of classifying brain tumor based on MRI images. The performance of the models is determined using the classification accuracy where VGG19 has the highest accuracy of 97.00 and ResNet50 has the second highest accuracy of 96.37. Inception3 and ResNet101 had the accuracies of 93.80 and 91.74, respectively. When comparing ResNet50 and VGG19, the deepest structures are found to capture more complicated features, and this makes them best in the classification of brain tumors. These results highlight the significance of proper choice of architecture in deep-learning applications to medical image analysis to improve automated diagnostic systems.