Scholarlink

<p>Lung cancer is one of the most common types of cancer worldwide and is responsible for a significant number of cancer-related deaths. Computer-aided diagnosis of lung images helps radiologists to treat the disease effectively. Histopathological image analysis, which is essentially microscopic imaging of lung tissues, is an important diagnostic method for lung cancer. In this paper, we develop a stacking model of deep learning (DL) based architectures to classify lung cancer using histopathological images. Initially, we extract features from three pre-trained NasNetLarge models (with the same architecture), each separately fed with color-transformed histopathological lung images, namely Red-Green-Blue (RGB), Cyan-Magenta-Yellow-Key (CMYK), and Luma-Blue minus Luma-Red minus Luma (YCbCr). Then we form a stacked multi-dimensional feature vector as obtained from the three DL models. This feature vector is then used to train a classification model using k-nearest neighbors (KNN). On a publicly available dataset of lung histopathological images, the proposed method achieves a high classification accuracy of 99.80%, for 3-class classification of lung cancer images. The source code of this work is available at <a href="https://github.com/Manish-Hazra/LCD">https://github.com/Manish-Hazra/LCD</a></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 stacking ensemble method with multiple color models for lung cancer classification using histopathology images

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Manish Hazra,
  • Nandita Gautam,
  • Jaydip Dey,
  • Ram Sarkar

摘要

Lung cancer is one of the most common types of cancer worldwide and is responsible for a significant number of cancer-related deaths. Computer-aided diagnosis of lung images helps radiologists to treat the disease effectively. Histopathological image analysis, which is essentially microscopic imaging of lung tissues, is an important diagnostic method for lung cancer. In this paper, we develop a stacking model of deep learning (DL) based architectures to classify lung cancer using histopathological images. Initially, we extract features from three pre-trained NasNetLarge models (with the same architecture), each separately fed with color-transformed histopathological lung images, namely Red-Green-Blue (RGB), Cyan-Magenta-Yellow-Key (CMYK), and Luma-Blue minus Luma-Red minus Luma (YCbCr). Then we form a stacked multi-dimensional feature vector as obtained from the three DL models. This feature vector is then used to train a classification model using k-nearest neighbors (KNN). On a publicly available dataset of lung histopathological images, the proposed method achieves a high classification accuracy of 99.80%, for 3-class classification of lung cancer images. The source code of this work is available at https://github.com/Manish-Hazra/LCD