M2S2L: Mamba-based Multi-Scale Spatial-temporal Learning for Video Anomaly Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Yang, Chen, Boan, Zhu, Xiaoguang, Liu, Jing, Sun, Peng, Zhou, Wei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912695482056704
author Liu, Yang
Chen, Boan
Zhu, Xiaoguang
Liu, Jing
Sun, Peng
Zhou, Wei
author_facet Liu, Yang
Chen, Boan
Zhu, Xiaoguang
Liu, Jing
Sun, Peng
Zhou, Wei
contents Video anomaly detection (VAD) is an essential task in the image processing community with prospects in video surveillance, which faces fundamental challenges in balancing detection accuracy with computational efficiency. As video content becomes increasingly complex with diverse behavioral patterns and contextual scenarios, traditional VAD approaches struggle to provide robust assessment for modern surveillance systems. Existing methods either lack comprehensive spatial-temporal modeling or require excessive computational resources for real-time applications. In this regard, we present a Mamba-based multi-scale spatial-temporal learning (M2S2L) framework in this paper. The proposed method employs hierarchical spatial encoders operating at multiple granularities and multi-temporal encoders capturing motion dynamics across different time scales. We also introduce a feature decomposition mechanism to enable task-specific optimization for appearance and motion reconstruction, facilitating more nuanced behavioral modeling and quality-aware anomaly assessment. Experiments on three benchmark datasets demonstrate that M2S2L framework achieves 98.5%, 92.1%, and 77.9% frame-level AUCs on UCSD Ped2, CUHK Avenue, and ShanghaiTech respectively, while maintaining efficiency with 20.1G FLOPs and 45 FPS inference speed, making it suitable for practical surveillance deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2511_05564
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle M2S2L: Mamba-based Multi-Scale Spatial-temporal Learning for Video Anomaly Detection
Liu, Yang
Chen, Boan
Zhu, Xiaoguang
Liu, Jing
Sun, Peng
Zhou, Wei
Computer Vision and Pattern Recognition
Video anomaly detection (VAD) is an essential task in the image processing community with prospects in video surveillance, which faces fundamental challenges in balancing detection accuracy with computational efficiency. As video content becomes increasingly complex with diverse behavioral patterns and contextual scenarios, traditional VAD approaches struggle to provide robust assessment for modern surveillance systems. Existing methods either lack comprehensive spatial-temporal modeling or require excessive computational resources for real-time applications. In this regard, we present a Mamba-based multi-scale spatial-temporal learning (M2S2L) framework in this paper. The proposed method employs hierarchical spatial encoders operating at multiple granularities and multi-temporal encoders capturing motion dynamics across different time scales. We also introduce a feature decomposition mechanism to enable task-specific optimization for appearance and motion reconstruction, facilitating more nuanced behavioral modeling and quality-aware anomaly assessment. Experiments on three benchmark datasets demonstrate that M2S2L framework achieves 98.5%, 92.1%, and 77.9% frame-level AUCs on UCSD Ped2, CUHK Avenue, and ShanghaiTech respectively, while maintaining efficiency with 20.1G FLOPs and 45 FPS inference speed, making it suitable for practical surveillance deployment.
title M2S2L: Mamba-based Multi-Scale Spatial-temporal Learning for Video Anomaly Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.05564