Scholarlink

<p>Lung cancer remains one of the most life-threatening diseases, and its early diagnosis continues to pose significant challenges. Without timely detection and treatment, the disease is often fatal. Computed tomography (CT) imaging is the most effective diagnostic modality, offering crucial insights into lung nodules, while histopathological analysis of Whole Slide Images (WSIs) provides confirmatory evidence. To enhance diagnostic accuracy and efficiency, this study proposes a novel cock–hen–chicken–vulture optimization (CHC–VO) algorithm integrated with a deep convolutional neural network (DCNN) for lung cancer detection. The CHC–VO algorithm optimizes the DCNN by adaptively selecting optimal weights and hyperparameters, thereby improving feature extraction and accelerating convergence. Lung lobe and cell segmentation are performed using the parallel reverse attention network (PRANET)<b>,</b> which refines boundary features through reverse attention and edge enhancement. Comparative analyses were conducted based on training data variations and k-fold validation. Experimental evaluation revealed that the CHC–VO–optimized DCNN attained remarkable performance, with an accuracy of 96.96%, sensitivity of 96.39%, specificity of 99.60%, precision of 95.81%, F1-score of 96.09%, and a Kappa coefficient of 97.35%<b>,</b> while maintaining a minimal FPR of 0.39<b>,</b> FNR of 3.61, and error rate of 3.035 using CT images. </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

CHC–VO enhanced PRANET for lung cancer detection using CT and histopathological images

  • 期刊论文

  • 发表日期：

    2026年1月4日

查看全文
  • P. S. Jeba,
  • R. S. Vinod Kumar,
  • S. S. Kumar,
  • D. Shahi

摘要

Lung cancer remains one of the most life-threatening diseases, and its early diagnosis continues to pose significant challenges. Without timely detection and treatment, the disease is often fatal. Computed tomography (CT) imaging is the most effective diagnostic modality, offering crucial insights into lung nodules, while histopathological analysis of Whole Slide Images (WSIs) provides confirmatory evidence. To enhance diagnostic accuracy and efficiency, this study proposes a novel cock–hen–chicken–vulture optimization (CHC–VO) algorithm integrated with a deep convolutional neural network (DCNN) for lung cancer detection. The CHC–VO algorithm optimizes the DCNN by adaptively selecting optimal weights and hyperparameters, thereby improving feature extraction and accelerating convergence. Lung lobe and cell segmentation are performed using the parallel reverse attention network (PRANET), which refines boundary features through reverse attention and edge enhancement. Comparative analyses were conducted based on training data variations and k-fold validation. Experimental evaluation revealed that the CHC–VO–optimized DCNN attained remarkable performance, with an accuracy of 96.96%, sensitivity of 96.39%, specificity of 99.60%, precision of 95.81%, F1-score of 96.09%, and a Kappa coefficient of 97.35%, while maintaining a minimal FPR of 0.39, FNR of 3.61, and error rate of 3.035 using CT images.