Efficient COVID-19 Detection using Optimized MobileNetV3-Small with SRGAN for Web Application

Abstract

Rapid and accurate detection of COVID-19 from medical images, such as X-rays and CT scans, is critical for timely diagnosis and treatment. This paper presents an innovative approach that combines Super-Resolution Generative Adversarial Network (SRGAN) for image enhancement with an optimized MobileNetV3-Small model to achieve efficient and high-accuracy classification. The proposed method significantly reduces computational complexity while maintaining performance. Specifically, the optimized MobileNetV3-Small model achieves 99.5% accuracy for X-ray images and 99.8% accuracy for CT images with only ~0.8M parameters and ~2.5 MB memory usage, making it highly suitable for real-time web applications in resource-constrained environments. Comparative analysis with related works demonstrates that the proposed approach outperforms other models in terms of accuracy, efficiency, and lightweight design. The results highlight the potential of the proposed method as a practical solution for rapid COVID-19 detection, contributing to the development of accessible and scalable diagnostic tools.