ACGAN for Addressing the Security Challenges in IoT-Based Healthcare System

Area of Science:

• Internet of Things (IoT)
• Cybersecurity in Healthcare
• Machine Learning

Background:

• IoT is widely used in healthcare for remote patient monitoring, enhancing care quality and patient lives.
• Healthcare Monitoring Systems (HMS) integrate devices and data for real-time patient management, but face significant data transmission security threats.
• Attacks on the Internet of Medical Things (IoMT) endanger patient safety due to data breaches.

Purpose of the Study:

• To address the challenge of scarce, imbalanced datasets for analyzing IoMT attack patterns.
• To enhance the security and integrity of data transmission in IoMT systems.
• To improve the detection of diverse IoMT cyber threats.

Main Methods:

• Utilized Auxiliary Classifier Generative Adversarial Network (ACGAN) to generate synthetic IoMT data samples.
• Incorporated biometric and network flow metrics for enhanced threat detection.
• Trained ACGAN with dual objectives for discriminator (real/fake classification and label prediction) and generator (high-quality, correctly labeled sample generation).

Main Results:

• The ACGAN approach generated high-quality synthetic data resembling minority class samples.
• Evaluated performance on the IoMT dataset, demonstrating superior results compared to baseline models.
• Achieved significant improvements across key metrics: accuracy, precision, recall, F1-score, AUC, and confusion matrix.

Conclusions:

• ACGAN effectively addresses the dataset scarcity issue in IoMT security research.
• The proposed method significantly enhances the detection of IoMT cyberattacks.
• This contributes to more secure and reliable remote patient monitoring and overall IoMT system integrity.