Scholarlink

The research paper addresses the pressing need for advanced methods in lung cancer detection and precise subtype identification to enhance patient care. Employing state-of-the-art convolutional neural network (CNN) architectures—Inception V4, VGG16, and ResNet—the study investigates their efficacy in detecting lung cancer from histopathological images and distinguishing between lung adenocarcinoma and squamous cell carcinoma. Key objectives include training these models using a comprehensive “Lung Cancer Preprocessed Dataset,” assessing their performance in accurate cancer detection and subtype classification, and conducting a comparative analysis to identify the most effective model. The exceptional performance of ResNet50, Inception V4, and VGG16 in fivefold cross-validation, achieving remarkable mean accuracies of 97%, 97%, 94% respectively, has profound implications for enhancing accurate classification of lung cancer.

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

Comparison of Accuracy in Lung Cancer Classification Through Convolutional Neural Network Models Based on Histopathological Image

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sajid Faysal Fahim,
  • Shodorson Nath,
  • Tanaj Afnan,
  • Md. Israkul Islam,
  • Tokiuddin Ahmed,
  • Sanjida Simla

摘要

The research paper addresses the pressing need for advanced methods in lung cancer detection and precise subtype identification to enhance patient care. Employing state-of-the-art convolutional neural network (CNN) architectures—Inception V4, VGG16, and ResNet—the study investigates their efficacy in detecting lung cancer from histopathological images and distinguishing between lung adenocarcinoma and squamous cell carcinoma. Key objectives include training these models using a comprehensive “Lung Cancer Preprocessed Dataset,” assessing their performance in accurate cancer detection and subtype classification, and conducting a comparative analysis to identify the most effective model. The exceptional performance of ResNet50, Inception V4, and VGG16 in fivefold cross-validation, achieving remarkable mean accuracies of 97%, 97%, 94% respectively, has profound implications for enhancing accurate classification of lung cancer.