Scholarlink

Brain cancer is a serious neurological condition that significantly impacts brain function and may be fatal. In India, it is estimated that 15 million cases are reported annually. It takes early discovery to lessen the tumor’s severity. However, manual examination of magnetic resonance imaging (MRI) data presents challenges, including inefficiencies and potential inaccuracies in diagnosing brain tumors. Brain tumors vary greatly in shape, size, appearance, and location, adding complexity to the diagnostic process. Computer-aided systems (CAD) have been developed for accurate and rapid diagnosis, but their efficacy still requires validation. The goal of this study is to identify the most efficient classification system among VGG16, VGG19, Inception-V3, and MobileNet-V2. Data augmentation methods are used to enhance the dataset before feeding it into these four convolutional neural networks (CNN) models. A comparison analysis is carried out with the aid of metrics like F1-score, recall, accuracy, and precision. The results show that VGG19 outperforms VGG16, MobileNet-V2, and Inception-V3 in terms of accuracy and precision when data augmentation is applied.

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

Strategic Enhancement of Brain Tumor Classification Using State-of-the-Art Transfer Learning Methods

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jyoti,
  • Anuj Kumar

摘要

Brain cancer is a serious neurological condition that significantly impacts brain function and may be fatal. In India, it is estimated that 15 million cases are reported annually. It takes early discovery to lessen the tumor’s severity. However, manual examination of magnetic resonance imaging (MRI) data presents challenges, including inefficiencies and potential inaccuracies in diagnosing brain tumors. Brain tumors vary greatly in shape, size, appearance, and location, adding complexity to the diagnostic process. Computer-aided systems (CAD) have been developed for accurate and rapid diagnosis, but their efficacy still requires validation. The goal of this study is to identify the most efficient classification system among VGG16, VGG19, Inception-V3, and MobileNet-V2. Data augmentation methods are used to enhance the dataset before feeding it into these four convolutional neural networks (CNN) models. A comparison analysis is carried out with the aid of metrics like F1-score, recall, accuracy, and precision. The results show that VGG19 outperforms VGG16, MobileNet-V2, and Inception-V3 in terms of accuracy and precision when data augmentation is applied.