Learning effective embedding for automated COVID-19 prediction from chest X-ray images

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic methods.
• Chest X-rays offer a more accessible alternative to complex tests like RT-PCR.
• Developing efficient deep learning models for COVID-19 detection from X-rays is crucial.

Purpose of the Study:

• To propose a novel deep learning model for unbiased COVID-19 detection from chest X-ray images.
• To enhance classifier performance and address class imbalance using transfer learning and data augmentation.
• To achieve state-of-the-art accuracy in COVID-19 detection.

Main Methods:

• Development of a novel convolution neural network (CNN) model.
• Utilizing transfer learning with pre-trained models (AlexNet, VGG16) on the ImageNet dataset.
• Implementation of data augmentation techniques to improve performance and handle class imbalance.

Main Results:

• The proposed CNN model achieved high accuracy and unbiased detection of COVID-19.
• The model outperformed existing deep learning models for chest X-ray-based COVID-19 detection.
• Demonstrated state-of-the-art performance with robust and scalable results.

Conclusions:

• The novel deep learning model provides an efficient and accurate method for COVID-19 detection using chest X-rays.
• The approach is easily deployable and scalable for widespread use.
• This research contributes to improved diagnostic efficiency in managing the COVID-19 pandemic.