Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Mumcu, Furkan, Jones, Michael J., Cherian, Anoop, Yilmaz, Yasin
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.12725
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917488591110144
author Mumcu, Furkan
Jones, Michael J.
Cherian, Anoop
Yilmaz, Yasin
author_facet Mumcu, Furkan
Jones, Michael J.
Cherian, Anoop
Yilmaz, Yasin
contents Recent video anomaly detection research has expanded rapidly with an emphasis on general models of normality intended to work across many different scenes. While this focus has led to improvements in scalability and multi-scene generalization, it has also shifted the field away from modeling the scene-specific and context-dependent nature of normal behavior. Contemporary approaches frequently rely on video-level weak supervision and opaque pretrained representations from multi-modal large language models (MLLMs), which encourage models to respond to familiar semantic anomaly categories rather than to deviations from the normal patterns of a particular environment. This trend suppresses spatial localization, introduces semantic bias, and reduces anomaly detection to a form of action recognition. In this paper, we examine whether these prevailing formulations align with the core requirements of real-world VAD, which is typically performed within a single scene where normality is determined by local geometry, semantics, and activity patterns. Through targeted visual analyses and empirical evaluations, we demonstrate the practical consequences of these limitations and show that meaningful progress in VAD requires renewed focus on single-scene, spatially-aware, and explainable formulations that capture the nuanced structure of normality within individual environments.
format Preprint
id arxiv_https___arxiv_org_abs_2605_12725
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Is Video Anomaly Detection Misframed? Evidence from LLM-Based and Multi-Scene Models
Mumcu, Furkan
Jones, Michael J.
Cherian, Anoop
Yilmaz, Yasin
Computer Vision and Pattern Recognition
Recent video anomaly detection research has expanded rapidly with an emphasis on general models of normality intended to work across many different scenes. While this focus has led to improvements in scalability and multi-scene generalization, it has also shifted the field away from modeling the scene-specific and context-dependent nature of normal behavior. Contemporary approaches frequently rely on video-level weak supervision and opaque pretrained representations from multi-modal large language models (MLLMs), which encourage models to respond to familiar semantic anomaly categories rather than to deviations from the normal patterns of a particular environment. This trend suppresses spatial localization, introduces semantic bias, and reduces anomaly detection to a form of action recognition. In this paper, we examine whether these prevailing formulations align with the core requirements of real-world VAD, which is typically performed within a single scene where normality is determined by local geometry, semantics, and activity patterns. Through targeted visual analyses and empirical evaluations, we demonstrate the practical consequences of these limitations and show that meaningful progress in VAD requires renewed focus on single-scene, spatially-aware, and explainable formulations that capture the nuanced structure of normality within individual environments.
title Is Video Anomaly Detection Misframed? Evidence from LLM-Based and Multi-Scene Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.12725