Related Experiment Video
Updated: Aug 10, 2025
04:17
DNA Virus Detection System Based on RPA-CRISPR/Cas12a-SPM and Deep Learning
Published on: May 10, 2024
840
Feature selection of pre-trained shallow CNN using the QLESCA optimizer: COVID-19 detection as a case study.
Qusay Shihab Hamad1,2, Hussein Samma3, Shahrel Azmin Suandi1
1Intelligent Biometric Group, School of Electrical and Electronic Engineering, Engineering Campus, Universiti Sains Malaysia, 14300 Nibong Tebal, Penang, Malaysia.
Applied Intelligence (Dordrecht, Netherlands)
|February 13, 2023
Summary
This study introduces a new method for detecting COVID-19 from X-rays using shallow CNNs and QLESCA for feature selection, achieving high accuracy. The approach enhances COVID-19 diagnosis efficiency by reducing computational costs and improving feature extraction.
Area of Science:
- Computer Science
- Medical Imaging
- Artificial Intelligence
Background:
- Millions of COVID-19 infections and deaths have occurred globally.
- Convolutional Neural Networks (CNNs) are used for COVID-19 detection from X-rays.
- Existing methods suffer from poor feature extraction and high computational costs, hindering accurate and rapid diagnosis.
Purpose of the Study:
- To develop an accurate, efficient, and computationally inexpensive method for COVID-19 feature extraction from chest X-rays.
- To address the 'curse of dimensionality' issue caused by poor feature extraction.
- To improve the performance of COVID-19 detection models.
Main Methods:
- A feature extraction mechanism based on Shallow Conventional Neural Network (SCNN).
- Feature selection using the Q-Learning Embedded Sine Cosine Algorithm (QLESCA).
- Classification using Support Vector Machines (SVM).
- Model trained and evaluated on five public chest X-ray datasets (4848 COVID-19, 8669 non-COVID-19 images).
Main Results:
- The proposed method achieved a high accuracy of 97.8086%.
- QLESCA reduced the number of features from 100 to 38.
- The model's accuracy improved with QLESCA-driven dimensionality reduction.
- QLESCA performance was superior to nine other optimization algorithms.
Conclusions:
- The developed approach offers an accurate and efficient solution for COVID-19 detection from chest X-rays.
- QLESCA is an effective technique for feature selection and dimensionality reduction in medical image analysis.
- This method can potentially overcome the limitations of existing CNN-based COVID-19 detection frameworks.
