COVID-19 detection from medical images using hummingbird-optimized VGG16 framework

Abstract

The rapid spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) created a major diagnostic challenge worldwide, highlighting the need for reliable automated approaches to assist clinicians in detecting coronavirus disease 2019 (COVID-19) from medical images. This study introduces a unified detection framework designed to deliver robust and accurate classification performance across different imaging modalities. The proposed model was developed and evaluated on computed tomography (CT) and chest X-ray datasets. The workflow comprises five key stages: image acquisition and resizing, deep-feature extraction, handcrafted feature extraction, feature optimization, and binary classification. The proposed system achieved high diagnostic performance, yielding approximately 99.33% accuracy for CT images and about 99.40% accuracy for X-ray images. The results demonstrate the framework’s robustness and its potential as effective tool for computer-aided COVID-19 diagnosis.