Related Experiment Video
Updated: Aug 27, 2025
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
14.3K
A COVID-19 X-ray image classification model based on an enhanced convolutional neural network and hill climbing
Ashwini Kumar Pradhan1, Debahuti Mishra1, Kaberi Das1
1Department of Computer Science and Engineering, Siksha O Anusandhan (Deemed to Be University), Khandagiri, Bhubaneswar, 751030 Odisha India.
Multimedia Tools and Applications
|October 3, 2022
Summary
A novel Convolution Neural Network with Hill-Climbing Algorithm (CNN-HCA) effectively classifies chest X-rays for COVID-19 and pneumonia. This deep learning approach demonstrates superior accuracy and performance over existing methods, aiding early disease detection.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Healthcare
- Deep Learning for Diagnostics
Background:
- Traditional medical image classification methods are time-consuming due to manual feature extraction.
- Deep learning, particularly Convolution Neural Networks (CNNs), offers high performance and self-learning capabilities for image analysis.
- Accurate and rapid classification of chest X-rays (CXRs) is crucial for diagnosing conditions like pneumonia and COVID-19.
Purpose of the Study:
- To evaluate and compare the performance of various pre-trained Convolution Neural Network (CNN) models for classifying chest X-ray (CXR) images as normal, pneumonia, or COVID-19.
- To propose and validate a novel hybrid CNN model, CNN-HCA, integrating the Hill-Climbing Algorithm for enhanced parameter optimization.
- To benchmark the proposed CNN-HCA model against existing hybrid classifiers and peer-reviewed studies using comprehensive performance metrics.
Main Methods:
- A comparative analysis was conducted using pre-trained CNN models including VGG-16, VGG-19, Inception version 3, Caps Net, DenseNet121, ResNet50, and Mobile-Net.
- A proposed CNN classifier was developed and evaluated.
- A novel CNN-HCA model was introduced, optimizing CNN parameters using the Hill-Climbing Algorithm, and compared with CNN-PSO and CNN-Jaya.
Main Results:
- The proposed CNN model showed potentially superior accuracy compared to other pre-trained models.
- The CNN-HCA model demonstrated superior performance across metrics such as Receiver Operating Characteristic Curve (ROC), Area Under the ROC Curve (AUC), sensitivity, specificity, F-score, and accuracy.
- Simulation findings indicated that CNN-HCA outperformed existing hybrid approaches in CXR classification.
Conclusions:
- The developed CNN-HCA model offers a promising and effective deep learning solution for automated CXR classification.
- The integration of the Hill-Climbing Algorithm significantly enhances CNN performance for medical image analysis.
- The study validates the potential of CNN-HCA for accurate and efficient early detection of respiratory illnesses from chest X-rays.
