Scholarlink

This research delves into the intricate landscape of Chronic Obstructive Pulmonary Disease (COPD) diagnosis, emphasizing the synergy between medical imaging, clinical knowledge, and advanced computational techniques. By leveraging tools like PyRadiomics and the CNN-ELM model, the study aims to enhance the precision of unveiling COPD patterns in High-Resolution Computed Tomography (HRCT) scans. The model training phase meticulously fine-tunes parameters to ensure accurate detection and classification of COPD, classified into phases according to GOLD stages: stage 0, stage I, stage II, stage III, stage IV, and stage V. The HRCT scans are labeled and grouped based on these classifications. The evaluation phase assesses the model using metrics such as accuracy, precision, recall, and F1-score, achieving notable accuracy (73.9%) and F1-score (0.239734) with 1600 hidden units and a test size of 40%. Additionally, the model’s AUC (Area under the ROC Curve) ranged from 0.5 to 0.9, with a peak of 0.93. These results underscore the model’s effectiveness in COPD detection, demonstrating improved performance compared to existing methods. This innovative approach promises a nuanced understanding of COPD, contributing to more personalized medical interventions through detailed insights and advanced diagnostics.

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

Integrated COPD Pattern Classification in HRCT Scans: PyRadiomics-CNN-ELM Hybrid Model Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Bianka Thomas,
  • Pon Harshavardhanan

摘要

This research delves into the intricate landscape of Chronic Obstructive Pulmonary Disease (COPD) diagnosis, emphasizing the synergy between medical imaging, clinical knowledge, and advanced computational techniques. By leveraging tools like PyRadiomics and the CNN-ELM model, the study aims to enhance the precision of unveiling COPD patterns in High-Resolution Computed Tomography (HRCT) scans. The model training phase meticulously fine-tunes parameters to ensure accurate detection and classification of COPD, classified into phases according to GOLD stages: stage 0, stage I, stage II, stage III, stage IV, and stage V. The HRCT scans are labeled and grouped based on these classifications. The evaluation phase assesses the model using metrics such as accuracy, precision, recall, and F1-score, achieving notable accuracy (73.9%) and F1-score (0.239734) with 1600 hidden units and a test size of 40%. Additionally, the model’s AUC (Area under the ROC Curve) ranged from 0.5 to 0.9, with a peak of 0.93. These results underscore the model’s effectiveness in COPD detection, demonstrating improved performance compared to existing methods. This innovative approach promises a nuanced understanding of COPD, contributing to more personalized medical interventions through detailed insights and advanced diagnostics.