Hybrid Model for DDOS Attack Detection and Covid-19 Classification
K. Nishitha Sree
Electronics and Communication Engineering Institute of Aeronautical Engineering Hyderabad, India nishithakeshaboina@iare.ac.in
Rufus Praleen
Electronics and Communication Engineering Institute of Aeronautical Engineering Hyderabad, India
21951a04f6@iare.ac.in
B. Naveen
Electronics and Communication Engineering Institute of Aeronautical Engineering
Hyderabad, India naveenbadavath670@gmail.com
Mr. G. Kiran Kumar
Electronics and Communication Engineering Institute of Aeronautical Engineering
Hyderabad, India g.kirankumar@iare.ac.in
Abstract
The COVID-19 pandemic has significantly impacted public INDEX TERMS: Chest X-ray image classification, health and economies worldwide. The integration of the convolutional neural networks, feistel block cipher, grey-wolf Internet of Things (IOT) with medical technology, known as optimization, Internet of Medical Things, hybrid ResNet50
the Internet of Medical Things (IOMT), has facilitated
advances like rapid diagnosis, remote monitoring, and more efficient healthcare delivery. However, the security, privacy,
and integrity of IOMT-generated medical data remain a critical challenge. Current systems are particularly vulnerable to cyber-attacks, such as Distributed Denial of Service (DDOS) attacks, which can disrupt medical services. To address these concerns, a new IOMT-based COVID-19 detection and classification system, named ICDC-Net, is proposed for smart healthcare applications. This system incorporates an Optimized Feistel Block Cipher (OFBC) for encryption to secure COVID-19-related medical data, particularly chest X-ray images, ensuring robust data protection. The OFBC algorithm is optimized using a hybrid approach combining the Gray Wolf Optimizer and Particle Swarm Optimization (HGWO-PSO), providing both encryption and effective DDOS attack detection and prevention. Additionally, the HGWO-PSO method is employed to extract features from chest X-rays, aiding in the detection of specific diseases. For disease classification, a deep learning model, Residual Network50 (ResNet50), is used to identify conditions like COVID-19, pneumonia, and other common lung diseases. Testing and simulations demonstrate that ICDC-Net enhances
