Scholarlink

<p>While considering the global health issue, early diagnosis of lung cancer has become a potential demand of any health-conscious system. Although the computer-aided diagnosis (CAD) system contributes a vital role in cancer-related disease they often fail to classify the different types of lung cancer due to computational complexity and high error rate. In this article, we have proposed a novel approach to classify lung cancer using transfer learning techniques and transformer-based models. The present work undertakes three major tasks: pre-processing, feature extractions, and classifications. Here, feature extraction is performed by transfer learning techniques using the VGG19 model, whereas the classification is performed with the help of the Vision Transformer (ViT) model. The combined approach of transfer learning and transformer-based model has been proven to classify lung cancer with the best accuracy of 99.80 and 99.06% over two datasets, LC25000 and IQ-OTH/NCCD, respectively. </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 Combined Approach of Vision Transformer and Transfer Learning Based Model for Accurate Lung Cancer Classification

  • 期刊论文

  • 发表日期：

    2025年5月30日

查看全文
  • Arvind Kumar,
  • Md Khalid Ansari,
  • Koushlendra Kumar Singh

摘要

While considering the global health issue, early diagnosis of lung cancer has become a potential demand of any health-conscious system. Although the computer-aided diagnosis (CAD) system contributes a vital role in cancer-related disease they often fail to classify the different types of lung cancer due to computational complexity and high error rate. In this article, we have proposed a novel approach to classify lung cancer using transfer learning techniques and transformer-based models. The present work undertakes three major tasks: pre-processing, feature extractions, and classifications. Here, feature extraction is performed by transfer learning techniques using the VGG19 model, whereas the classification is performed with the help of the Vision Transformer (ViT) model. The combined approach of transfer learning and transformer-based model has been proven to classify lung cancer with the best accuracy of 99.80 and 99.06% over two datasets, LC25000 and IQ-OTH/NCCD, respectively.