Scholarlink

<p>Smart Door security is emerging as the top concern for people around the world. In the past decade, numerous strategies have been utilized including door locks, bio-metric authentication to limit the intrusion of unauthorized individuals. To overcome these prior challenges facial recognition system has been adopted. However, existing studies demonstrate that facial recognition systems can be spoofed by displaying highresolution videos of the person. The motivation of this study is to develop an appropriate explainable AI smart door security employing face localization &amp; identification incorporating PIR sensors based on transfer learning. The PIR sensor helps to validate the presence of the actual person making it different &amp; more reliable from earlier approaches. To achieve our objective, a new data set have been developed including people with long beard. The methodology not only helps limit intrusion but also plays an important role in mitigating security concerns for individuals and families.</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

An explainable AI-enhanced smart door security system using transfer learning

  • 期刊论文

  • 发表日期：

    2025年6月30日

查看全文
  • Murtaza Hanif,
  • Taj Muhammad,
  • Muhammad Junaid Arshad

摘要

Smart Door security is emerging as the top concern for people around the world. In the past decade, numerous strategies have been utilized including door locks, bio-metric authentication to limit the intrusion of unauthorized individuals. To overcome these prior challenges facial recognition system has been adopted. However, existing studies demonstrate that facial recognition systems can be spoofed by displaying highresolution videos of the person. The motivation of this study is to develop an appropriate explainable AI smart door security employing face localization & identification incorporating PIR sensors based on transfer learning. The PIR sensor helps to validate the presence of the actual person making it different & more reliable from earlier approaches. To achieve our objective, a new data set have been developed including people with long beard. The methodology not only helps limit intrusion but also plays an important role in mitigating security concerns for individuals and families.