Research on Classification of COVID-19 Chest X-Ray Image Modal Feature Fusion Based on Deep Learning

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Infectious Disease Diagnostics

Background:

• Current COVID-19 detection methods using chest X-rays often rely on classic image classification models.
• These traditional models face challenges with low recognition accuracy and inadequate capture of crucial modal features.
• There is a need for enhanced diagnostic tools to improve the accuracy of COVID-19 detection from radiographic images.

Purpose of the Study:

• To propose and evaluate a novel COVID-19 detection method based on image modal feature fusion.
• To enhance the accuracy and reliability of COVID-19 diagnosis using chest X-ray imaging.
• To address the limitations of existing classification models in capturing essential image features for COVID-19.

Main Methods:

• Chest X-rays underwent small-sample enhancement processing, including rotation, translation, and random transformations.
• Modal features were extracted using five classic pre-trained models.
• A global average pooling layer was employed to reduce parameters and prevent overfitting, followed by model training and fine-tuning.

Main Results:

• The proposed method demonstrated superior performance in detecting COVID-19 image modal information compared to classic models.
• Evaluation using machine learning standards and receiver operating characteristic (ROC) curves confirmed the effectiveness of the approach.
• The method achieved the expected outcome of accurately detecting COVID-19 cases from chest X-rays.

Conclusions:

• The image modal feature fusion method offers a more effective approach for COVID-19 detection from chest X-rays.
• This technique improves upon traditional models by better capturing and utilizing image features for diagnosis.
• The study validates the potential of this enhanced method for accurate and reliable COVID-19 case detection.