Scholarlink

<p>One of the key features emerging within the smart healthcare and Internet of things framework is the increasing necessity for effective intrusion detection systems, particularly as cybersecurity attacks and privacy concerns become more prevalent in medical environments. The protection of sensitive patient data and integrity of healthcare systems are critical, making it imperative to develop robust solutions that can identify and mitigate potential threats in real time. This paper proposes an adaptive generative vision transformer network specifically designed for accurate and early intrusion detection in smart healthcare systems. The proposed adaptive generative vision transformer network effectively diagnoses and categorizes data from widely used intrusion detection datasets. The methodology involves several stages, starting with data discovery, followed by preprocessing steps that include dataset collection, normalization, and data cleaning. Using SHapley Additive exPlanations, the model evaluates key characteristics such as originality and stability, enabling the identification of the most influential features for intrusion detection. In addition, the incremental wrapper subset selection with replacement method is employed to eliminate redundant and irrelevant features, improving computational efficiency and classification accuracy.Subsequently, fine-grained classification of raw data is achieved through probability estimates, distinguishing between normal and abnormal categories. This integration enhances the vision transformer with generative adversarial networks, creating a robust solution for the early detection of cyber threats, ultimately ensuring the security of smart healthcare environments. Performance analysis of the proposed model reveals a high accuracy and detection rate. This study explicits the necessity and relevance of improving the functionality of intrusion detection systems and helps improve privacy and authentication in smart healthcare systems. Experimental results demonstrate up to 99.2% accuracy and improved detection rates, validating the effectiveness of AG-ViTNet in securing smart healthcare network.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advanced intrusion detection in internet of things-driven health care with adaptive generative vision transformer network and generative vision transformers

  • 期刊论文

  • 发表日期：

    2025年9月16日

查看全文
  • T. Thiyagu,
  • S. Krishnaveni

摘要

One of the key features emerging within the smart healthcare and Internet of things framework is the increasing necessity for effective intrusion detection systems, particularly as cybersecurity attacks and privacy concerns become more prevalent in medical environments. The protection of sensitive patient data and integrity of healthcare systems are critical, making it imperative to develop robust solutions that can identify and mitigate potential threats in real time. This paper proposes an adaptive generative vision transformer network specifically designed for accurate and early intrusion detection in smart healthcare systems. The proposed adaptive generative vision transformer network effectively diagnoses and categorizes data from widely used intrusion detection datasets. The methodology involves several stages, starting with data discovery, followed by preprocessing steps that include dataset collection, normalization, and data cleaning. Using SHapley Additive exPlanations, the model evaluates key characteristics such as originality and stability, enabling the identification of the most influential features for intrusion detection. In addition, the incremental wrapper subset selection with replacement method is employed to eliminate redundant and irrelevant features, improving computational efficiency and classification accuracy.Subsequently, fine-grained classification of raw data is achieved through probability estimates, distinguishing between normal and abnormal categories. This integration enhances the vision transformer with generative adversarial networks, creating a robust solution for the early detection of cyber threats, ultimately ensuring the security of smart healthcare environments. Performance analysis of the proposed model reveals a high accuracy and detection rate. This study explicits the necessity and relevance of improving the functionality of intrusion detection systems and helps improve privacy and authentication in smart healthcare systems. Experimental results demonstrate up to 99.2% accuracy and improved detection rates, validating the effectiveness of AG-ViTNet in securing smart healthcare network.