Short-Window Sliding Learning for Real-Time Violence Detection via LLM-based Auto-Labeling

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jung, Seoik, Song, Taekyung, Lee, Yangro, Lee, Sungjun
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909901073154048
author Jung, Seoik
Song, Taekyung
Lee, Yangro
Lee, Sungjun
author_facet Jung, Seoik
Song, Taekyung
Lee, Yangro
Lee, Sungjun
contents This paper proposes a Short-Window Sliding Learning framework for real-time violence detection in CCTV footages. Unlike conventional long-video training approaches, the proposed method divides videos into 1-2 second clips and applies Large Language Model (LLM)-based auto-caption labeling to construct fine-grained datasets. Each short clip fully utilizes all frames to preserve temporal continuity, enabling precise recognition of rapid violent events. Experiments demonstrate that the proposed method achieves 95.25\% accuracy on RWF-2000 and significantly improves performance on long videos (UCF-Crime: 83.25\%), confirming its strong generalization and real-time applicability in intelligent surveillance systems.
format Preprint
id arxiv_https___arxiv_org_abs_2511_10866
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Short-Window Sliding Learning for Real-Time Violence Detection via LLM-based Auto-Labeling
Jung, Seoik
Song, Taekyung
Lee, Yangro
Lee, Sungjun
Computer Vision and Pattern Recognition
Artificial Intelligence
68T45, 68T07
I.2.10; I.4.8; I.2.6
This paper proposes a Short-Window Sliding Learning framework for real-time violence detection in CCTV footages. Unlike conventional long-video training approaches, the proposed method divides videos into 1-2 second clips and applies Large Language Model (LLM)-based auto-caption labeling to construct fine-grained datasets. Each short clip fully utilizes all frames to preserve temporal continuity, enabling precise recognition of rapid violent events. Experiments demonstrate that the proposed method achieves 95.25\% accuracy on RWF-2000 and significantly improves performance on long videos (UCF-Crime: 83.25\%), confirming its strong generalization and real-time applicability in intelligent surveillance systems.
title Short-Window Sliding Learning for Real-Time Violence Detection via LLM-based Auto-Labeling
topic Computer Vision and Pattern Recognition
Artificial Intelligence
68T45, 68T07
I.2.10; I.4.8; I.2.6
url https://arxiv.org/abs/2511.10866