Scholarlink

<p>The automatic recognition of human behavior is attracting more research attention, especially with the rapid progress of neural networks in recent years. The efficient detection and recognition of abnormal human behavior (AHB) are critical components of intelligent video surveillance systems, as they ensure human security and create safe environments. This paper presents a systematic literature review (SLR) of 140 studies published between 2016 and 2024, focusing on the application of deep learning for AHB recognition in videos. We formulate eight key research questions (RQs) that explore the types of AHB addressed in the literature, the historical progression of deep learning models, pre-trained architectures, existing methods, real-time applications, datasets, performance metrics, and future research directions. This review paper serves as a valuable guide for both academia and industry professionals seeking to understand this field and explore the top emerging trends.</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

Deep learning applied for abnormal human behavior recognition in video surveillance systems: A systematic review

  • 期刊论文

  • 发表日期：

    2025年8月2日

查看全文
  • Olfa Saket,
  • Anis Ben Aicha,
  • Habib Fathallah

摘要

The automatic recognition of human behavior is attracting more research attention, especially with the rapid progress of neural networks in recent years. The efficient detection and recognition of abnormal human behavior (AHB) are critical components of intelligent video surveillance systems, as they ensure human security and create safe environments. This paper presents a systematic literature review (SLR) of 140 studies published between 2016 and 2024, focusing on the application of deep learning for AHB recognition in videos. We formulate eight key research questions (RQs) that explore the types of AHB addressed in the literature, the historical progression of deep learning models, pre-trained architectures, existing methods, real-time applications, datasets, performance metrics, and future research directions. This review paper serves as a valuable guide for both academia and industry professionals seeking to understand this field and explore the top emerging trends.