A Survey on Video Anomaly Detection via Deep Learning: Human, Vehicle, and Environment

2025-08-19

AI Summary

Video anomaly detection (VAD) faces core challenges including poor cross-domain generalization, fragmented learning paradigms, and difficulties in real-world deployment. This paper presents the first systematic survey of deep learning–based VAD, categorizing approaches by human-centric, vehicle-centric, and environment-centric application scenarios, and unifying supervised, weakly supervised, unsupervised, online, active, and continual learning paradigms. Through horizontal comparative analysis, it identifies shared limitations across methods in modeling capacity, annotation efficiency, and dynamic adaptability. The work establishes the first structured knowledge framework covering multiple supervision levels and adaptive learning paradigms. Furthermore, it distills six open challenges for practical deployment: foundational modeling, evaluation benchmarks, long-tailed anomalies, computational efficiency, model interpretability, and system robustness. This survey serves as a synergistic research roadmap bridging theoretical advancement and industrial adoption in VAD.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Video Anomaly Detection (VAD) has emerged as a pivotal task in computer vision, with broad relevance across multiple fields. Recent advances in deep learning have driven significant progress in this area, yet the field remains fragmented across domains and learning paradigms. This survey offers a comprehensive perspective on VAD, systematically organizing the literature across various supervision levels, as well as adaptive learning methods such as online, active, and continual learning. We examine the state of VAD across three major application categories: human-centric, vehicle-centric, and environment-centric scenarios, each with distinct challenges and design considerations. In doing so, we identify fundamental contributions and limitations of current methodologies. By consolidating insights from subfields, we aim to provide the community with a structured foundation for advancing both theoretical understanding and real-world applicability of VAD systems. This survey aims to support researchers by providing a useful reference, while also drawing attention to the broader set of open challenges in anomaly detection, including both fundamental research questions and practical obstacles to real-world deployment.

Problem

• Surveying deep learning approaches for video anomaly detection
• Organizing fragmented literature across supervision levels and applications
• Addressing challenges in human, vehicle, and environment scenarios

Innovation

• Deep learning for video anomaly detection
• Survey across supervision levels and learning
• Application categories: human, vehicle, environment