Flashback: Memory-Driven Zero-shot, Real-time Video Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Hyogun, Kim, Haksub, Kim, Ig-Jae, Choi, Yonghun
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913854515052544
author Lee, Hyogun
Kim, Haksub
Kim, Ig-Jae
Choi, Yonghun
author_facet Lee, Hyogun
Kim, Haksub
Kim, Ig-Jae
Choi, Yonghun
contents Video Anomaly Detection (VAD) automatically identifies anomalous events from video, mitigating the need for human operators in large-scale surveillance deployments. However, two fundamental obstacles hinder real-world adoption: domain dependency and real-time constraints -- requiring near-instantaneous processing of incoming video. To this end, we propose Flashback, a zero-shot and real-time video anomaly detection paradigm. Inspired by the human cognitive mechanism of instantly judging anomalies and reasoning in current scenes based on past experience, Flashback operates in two stages: Recall and Respond. In the offline recall stage, an off-the-shelf LLM builds a pseudo-scene memory of both normal and anomalous captions without any reliance on real anomaly data. In the online respond stage, incoming video segments are embedded and matched against this memory via similarity search. By eliminating all LLM calls at inference time, Flashback delivers real-time VAD even on a consumer-grade GPU. On two large datasets from real-world surveillance scenarios, UCF-Crime and XD-Violence, we achieve 87.3 AUC (+7.0 pp) and 75.1 AP (+13.1 pp), respectively, outperforming prior zero-shot VAD methods by large margins.
format Preprint
id arxiv_https___arxiv_org_abs_2505_15205
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Flashback: Memory-Driven Zero-shot, Real-time Video Anomaly Detection
Lee, Hyogun
Kim, Haksub
Kim, Ig-Jae
Choi, Yonghun
Computer Vision and Pattern Recognition
Video Anomaly Detection (VAD) automatically identifies anomalous events from video, mitigating the need for human operators in large-scale surveillance deployments. However, two fundamental obstacles hinder real-world adoption: domain dependency and real-time constraints -- requiring near-instantaneous processing of incoming video. To this end, we propose Flashback, a zero-shot and real-time video anomaly detection paradigm. Inspired by the human cognitive mechanism of instantly judging anomalies and reasoning in current scenes based on past experience, Flashback operates in two stages: Recall and Respond. In the offline recall stage, an off-the-shelf LLM builds a pseudo-scene memory of both normal and anomalous captions without any reliance on real anomaly data. In the online respond stage, incoming video segments are embedded and matched against this memory via similarity search. By eliminating all LLM calls at inference time, Flashback delivers real-time VAD even on a consumer-grade GPU. On two large datasets from real-world surveillance scenarios, UCF-Crime and XD-Violence, we achieve 87.3 AUC (+7.0 pp) and 75.1 AP (+13.1 pp), respectively, outperforming prior zero-shot VAD methods by large margins.
title Flashback: Memory-Driven Zero-shot, Real-time Video Anomaly Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.15205