Scholarlink

This study aims to improve the classification of lung and colon cancer types in histopathological images by simultaneously using deep learning feature extraction alongside advanced denoising algorithms. It evaluates traditional handcrafted feature extraction methods, as well as deep learning-based models, such as MobileNet, Xception, and a hybrid architecture combining MobileNet-Xception. Additional experiments are conducted with and without denoising. Baseline denoising methods, like Gaussian and wavelet denoising, as well as deep learning architectures such as U-Net and DnCNN, are employed. The U-Net denoising model, paired with the hybrid MobileNet-Xception model, achieved a high accuracy of 98.7%.

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

Enhancing Lung and Colon Disease Classification in Histopathological Images Using Denoising and Computationally Efficient Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • K. S. H. Pranathi,
  • G. Kalyani,
  • B. Kalyani

摘要

This study aims to improve the classification of lung and colon cancer types in histopathological images by simultaneously using deep learning feature extraction alongside advanced denoising algorithms. It evaluates traditional handcrafted feature extraction methods, as well as deep learning-based models, such as MobileNet, Xception, and a hybrid architecture combining MobileNet-Xception. Additional experiments are conducted with and without denoising. Baseline denoising methods, like Gaussian and wavelet denoising, as well as deep learning architectures such as U-Net and DnCNN, are employed. The U-Net denoising model, paired with the hybrid MobileNet-Xception model, achieved a high accuracy of 98.7%.