Scholarlink

<p>Brain tumors are one of the fastest growing cancers in the world, Accurate and early diagnosis of brain tumors is crucial to save patients’ lives. In order to rectify these problems, this work proposes a Cross Attention Transformer based Dragonfly Optimized Kernel Extreme Learning Machine method for accurate brain tumor detection and classification. A K-means clustering algorithm is utilized for the detection of brain tumors from enhanced images. Feature extraction is mainly used to improve the performance and accuracy of the proposed model by reducing the amount of resources used without losing any important information from the image to be identified. Then they discriminate the tumor-free image from those images. The Dragonfly Optimized Kernel Extreme Learning Machine method is used to classify the tumors. The kernel-based Extreme Learning Machine model is developed to improve the classification accuracy of ELM, and Dragonfly Optimization Algorithm method is used for classification. For the identification of lesions in the brain images, the K-means clustering method is applied and the generator and discriminator model are used to generate images using a cross-attention transducer-based autoencoder and discriminator. Dragonfly Optimized Kernel Extreme Learning Machine method is also used to better classify the tumor. The proposed method is evaluated on various brain tumor related datasets that proved the best performance of the proposed method against traditional approaches. The highest classification accuracy of 98.89% is attained for the proposed method during the comprehensive analysis that helps doctors to diagnose brain tumors accurately.</p>

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

A novel machine learning framework: cross transformer based optimization model for the detection and classification of brain tumor using clinical decision analysis

  • 期刊论文

  • 发表日期：

    2025年7月9日

查看全文
  • A. V. R. Mayuri,
  • S. P. Maniraj,
  • M. Duraisamy,
  • G. L. N. Murthy,
  • Kanika Garg,
  • M. Sangeetha

摘要

Brain tumors are one of the fastest growing cancers in the world, Accurate and early diagnosis of brain tumors is crucial to save patients’ lives. In order to rectify these problems, this work proposes a Cross Attention Transformer based Dragonfly Optimized Kernel Extreme Learning Machine method for accurate brain tumor detection and classification. A K-means clustering algorithm is utilized for the detection of brain tumors from enhanced images. Feature extraction is mainly used to improve the performance and accuracy of the proposed model by reducing the amount of resources used without losing any important information from the image to be identified. Then they discriminate the tumor-free image from those images. The Dragonfly Optimized Kernel Extreme Learning Machine method is used to classify the tumors. The kernel-based Extreme Learning Machine model is developed to improve the classification accuracy of ELM, and Dragonfly Optimization Algorithm method is used for classification. For the identification of lesions in the brain images, the K-means clustering method is applied and the generator and discriminator model are used to generate images using a cross-attention transducer-based autoencoder and discriminator. Dragonfly Optimized Kernel Extreme Learning Machine method is also used to better classify the tumor. The proposed method is evaluated on various brain tumor related datasets that proved the best performance of the proposed method against traditional approaches. The highest classification accuracy of 98.89% is attained for the proposed method during the comprehensive analysis that helps doctors to diagnose brain tumors accurately.