Related Experiment Video
Updated: Aug 15, 2025
04:17
DNA Virus Detection System Based on RPA-CRISPR/Cas12a-SPM and Deep Learning
Published on: May 10, 2024
842
Development of CNN-LSTM combinational architecture for COVID-19 detection
Abhinav Narula1, Naveen Kumar Vaegae1
1School of Electronics Engineering, Vellore Institute of Technology, Vellore, 632014 India.
Journal of Ambient Intelligence and Humanized Computing
|January 2, 2023
Summary
A novel deep learning model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks accurately detects COVID-19 from chest X-rays. This approach offers a reliable alternative to PCR tests, achieving high accuracy in classifying viral pneumonia.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computational Biology
Background:
- The global COVID-19 pandemic necessitated rapid and accurate diagnostic tools.
- Limitations in Reverse Transcription-Polymerase Chain Reaction (RT-PCR) testing, including anomalies in results, highlighted the need for alternative diagnostic methods.
- Computed Tomography (CT) scan-based and X-ray imaging analyses emerged as promising avenues for COVID-19 detection.
Purpose of the Study:
- To develop and validate a novel deep learning model for accurate COVID-19 detection using chest X-ray images.
- To leverage a combinational architecture of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for enhanced classification performance.
- To address the limitations of existing diagnostic methods by providing a more reliable and informative classification approach.
Main Methods:
- A hybrid CNN-LSTM deep learning architecture was designed, incorporating 8 convolutional and 4 pooling layers for feature extraction (CNN) and 4 LSTM layers for classification.
- Rectified Linear Unit (ReLU) activation function was utilized to introduce non-linearity and improve accuracy.
- A padding layer was implemented to prevent information loss during processing, and the model was trained and validated on a dataset of 7292 chest X-ray images.
Main Results:
- The proposed CNN-LSTM model achieved a high accuracy of 98.91% in classifying COVID-19 cases from chest X-ray images.
- A Matthew's Correlation Coefficient (MCC) of 97.84% was obtained, indicating excellent agreement between predicted and actual classifications.
- Performance metrics including accuracy, loss, F1 score, and MCC were rigorously calculated to evaluate the model's effectiveness.
Conclusions:
- The developed CNN-LSTM model demonstrates superior performance in COVID-19 detection compared to traditional methods and transfer learning approaches.
- The combinational architecture effectively extracts relevant features and classifies samples, providing informative and truthful results.
- This deep learning approach offers a promising, accurate, and reliable tool for augmenting COVID-19 diagnosis, particularly when RT-PCR results are uncertain.
