Scholarlink

<p>The field of medical image analysis has been transformed by deep learning, which offers robust methods for disease diagnosis. Optimization algorithms further enhance performance by selecting the most relevant features, therefore improving diagnosis accuracy. This study presents an enhanced method for classifying lung diseases, including COVID-19, viral pneumonia, and non-COVID conditions, using chest X-rays and CT images. Deep features are extracted from pretrained DenseNet-121 and ResNet-50 models and fused into a single, comprehensive feature set. To optimize feature selection, we apply the Honey Badger Algorithm (HBA). HBA effectively reduces the dimensionality of fused features while preserving important information. The selected features are used to train a Support Vector Machine (SVM). The robustness of the approach is demonstrated by the high classification accuracy and substantial gains across performance metrics observed in the experimental results. Overall, the diagnostic accuracy of medical image analysis is improved by the integration of feature fusion with HBA based selection.</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

Deep feature fusion and selection for lung disease classification

  • 期刊论文

  • 发表日期：

    2026年2月7日

查看全文
  • Kevisino Khate,
  • Arambam Neelima

摘要

The field of medical image analysis has been transformed by deep learning, which offers robust methods for disease diagnosis. Optimization algorithms further enhance performance by selecting the most relevant features, therefore improving diagnosis accuracy. This study presents an enhanced method for classifying lung diseases, including COVID-19, viral pneumonia, and non-COVID conditions, using chest X-rays and CT images. Deep features are extracted from pretrained DenseNet-121 and ResNet-50 models and fused into a single, comprehensive feature set. To optimize feature selection, we apply the Honey Badger Algorithm (HBA). HBA effectively reduces the dimensionality of fused features while preserving important information. The selected features are used to train a Support Vector Machine (SVM). The robustness of the approach is demonstrated by the high classification accuracy and substantial gains across performance metrics observed in the experimental results. Overall, the diagnostic accuracy of medical image analysis is improved by the integration of feature fusion with HBA based selection.