DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection

Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN an...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Ruituo, Chen, Yang, Xiao, Jian, Li, Bing, Fan, Jicong, Dufaux, Frédéric, Zhu, Ce, Liu, Yipeng
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Wu, Ruituo
Chen, Yang
Xiao, Jian
Li, Bing
Fan, Jicong
Dufaux, Frédéric
Zhu, Ce
Liu, Yipeng
description Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN and TCN blocks are constrained by small receptive fields and a lack of cross-dimension interaction capture. To tackle this limitation, we propose a lightweight module called the Dual Attention Module (DAM) for capturing cross-dimension interaction relationships in spatio-temporal skeletal data. It employs the frame attention mechanism to identify the most significant frames and the skeleton attention mechanism to capture broader relationships across fixed partitions with minimal parameters and flops. Furthermore, the proposed Dual Attention Normalizing Flow (DA-Flow) integrates the DAM as a post-processing unit after GCN within the normalizing flow framework. Simulations show that the proposed model is robust against noise and negative samples. Experimental results show that DA-Flow reaches competitive or better performance than the existing state-of-the-art (SOTA) methods in terms of the micro AUC metric with the fewest number of parameters. Moreover, we found that even without training, simply using random projection without dimensionality reduction on skeleton data enables substantial anomaly detection capabilities.
doi_str_mv 10.48550/arxiv.2406.02976
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2406_02976</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2406_02976</sourcerecordid><originalsourceid>FETCH-LOGICAL-a676-b7a8375578adebaf6efa4ce2136d5e53f80831b0cb2b9d3be95966c0510ac4743</originalsourceid><addsrcrecordid>eNotj7tOwzAYRr0woMIDMOEXcHDia9iihgJSBUMr1uh3_BtZpDFKzaU8PaR0-pbzHekQclXyQlql-A1M3_GzqCTXBa9qo8_Jpm3Yakhft7T9gIE2OeOYYxrpU5p2MMSfOL7SGaAhTXTzhgPmNDIHe_T0JXpMtBnTH3mgLWbs5-8FOQsw7PHytAuyXd1tlw9s_Xz_uGzWDLTRzBmwwihlLHh0EDQGkD1WpdBeoRLBcitKx3tXudoLh7Wqte65Kjn00kixINf_2mNV9z7FHUyHbq7rjnXiF-ZySmY</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection</title><source>arXiv.org</source><creator>Wu, Ruituo ; Chen, Yang ; Xiao, Jian ; Li, Bing ; Fan, Jicong ; Dufaux, Frédéric ; Zhu, Ce ; Liu, Yipeng</creator><creatorcontrib>Wu, Ruituo ; Chen, Yang ; Xiao, Jian ; Li, Bing ; Fan, Jicong ; Dufaux, Frédéric ; Zhu, Ce ; Liu, Yipeng</creatorcontrib><description>Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN and TCN blocks are constrained by small receptive fields and a lack of cross-dimension interaction capture. To tackle this limitation, we propose a lightweight module called the Dual Attention Module (DAM) for capturing cross-dimension interaction relationships in spatio-temporal skeletal data. It employs the frame attention mechanism to identify the most significant frames and the skeleton attention mechanism to capture broader relationships across fixed partitions with minimal parameters and flops. Furthermore, the proposed Dual Attention Normalizing Flow (DA-Flow) integrates the DAM as a post-processing unit after GCN within the normalizing flow framework. Simulations show that the proposed model is robust against noise and negative samples. Experimental results show that DA-Flow reaches competitive or better performance than the existing state-of-the-art (SOTA) methods in terms of the micro AUC metric with the fewest number of parameters. Moreover, we found that even without training, simply using random projection without dimensionality reduction on skeleton data enables substantial anomaly detection capabilities.</description><identifier>DOI: 10.48550/arxiv.2406.02976</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2024-06</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2406.02976$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2406.02976$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Wu, Ruituo</creatorcontrib><creatorcontrib>Chen, Yang</creatorcontrib><creatorcontrib>Xiao, Jian</creatorcontrib><creatorcontrib>Li, Bing</creatorcontrib><creatorcontrib>Fan, Jicong</creatorcontrib><creatorcontrib>Dufaux, Frédéric</creatorcontrib><creatorcontrib>Zhu, Ce</creatorcontrib><creatorcontrib>Liu, Yipeng</creatorcontrib><title>DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection</title><description>Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN and TCN blocks are constrained by small receptive fields and a lack of cross-dimension interaction capture. To tackle this limitation, we propose a lightweight module called the Dual Attention Module (DAM) for capturing cross-dimension interaction relationships in spatio-temporal skeletal data. It employs the frame attention mechanism to identify the most significant frames and the skeleton attention mechanism to capture broader relationships across fixed partitions with minimal parameters and flops. Furthermore, the proposed Dual Attention Normalizing Flow (DA-Flow) integrates the DAM as a post-processing unit after GCN within the normalizing flow framework. Simulations show that the proposed model is robust against noise and negative samples. Experimental results show that DA-Flow reaches competitive or better performance than the existing state-of-the-art (SOTA) methods in terms of the micro AUC metric with the fewest number of parameters. Moreover, we found that even without training, simply using random projection without dimensionality reduction on skeleton data enables substantial anomaly detection capabilities.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj7tOwzAYRr0woMIDMOEXcHDia9iihgJSBUMr1uh3_BtZpDFKzaU8PaR0-pbzHekQclXyQlql-A1M3_GzqCTXBa9qo8_Jpm3Yakhft7T9gIE2OeOYYxrpU5p2MMSfOL7SGaAhTXTzhgPmNDIHe_T0JXpMtBnTH3mgLWbs5-8FOQsw7PHytAuyXd1tlw9s_Xz_uGzWDLTRzBmwwihlLHh0EDQGkD1WpdBeoRLBcitKx3tXudoLh7Wqte65Kjn00kixINf_2mNV9z7FHUyHbq7rjnXiF-ZySmY</recordid><startdate>20240605</startdate><enddate>20240605</enddate><creator>Wu, Ruituo</creator><creator>Chen, Yang</creator><creator>Xiao, Jian</creator><creator>Li, Bing</creator><creator>Fan, Jicong</creator><creator>Dufaux, Frédéric</creator><creator>Zhu, Ce</creator><creator>Liu, Yipeng</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240605</creationdate><title>DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection</title><author>Wu, Ruituo ; Chen, Yang ; Xiao, Jian ; Li, Bing ; Fan, Jicong ; Dufaux, Frédéric ; Zhu, Ce ; Liu, Yipeng</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a676-b7a8375578adebaf6efa4ce2136d5e53f80831b0cb2b9d3be95966c0510ac4743</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Wu, Ruituo</creatorcontrib><creatorcontrib>Chen, Yang</creatorcontrib><creatorcontrib>Xiao, Jian</creatorcontrib><creatorcontrib>Li, Bing</creatorcontrib><creatorcontrib>Fan, Jicong</creatorcontrib><creatorcontrib>Dufaux, Frédéric</creatorcontrib><creatorcontrib>Zhu, Ce</creatorcontrib><creatorcontrib>Liu, Yipeng</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Wu, Ruituo</au><au>Chen, Yang</au><au>Xiao, Jian</au><au>Li, Bing</au><au>Fan, Jicong</au><au>Dufaux, Frédéric</au><au>Zhu, Ce</au><au>Liu, Yipeng</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection</atitle><date>2024-06-05</date><risdate>2024</risdate><abstract>Cooperation between temporal convolutional networks (TCN) and graph convolutional networks (GCN) as a processing module has shown promising results in skeleton-based video anomaly detection (SVAD). However, to maintain a lightweight model with low computational and storage complexity, shallow GCN and TCN blocks are constrained by small receptive fields and a lack of cross-dimension interaction capture. To tackle this limitation, we propose a lightweight module called the Dual Attention Module (DAM) for capturing cross-dimension interaction relationships in spatio-temporal skeletal data. It employs the frame attention mechanism to identify the most significant frames and the skeleton attention mechanism to capture broader relationships across fixed partitions with minimal parameters and flops. Furthermore, the proposed Dual Attention Normalizing Flow (DA-Flow) integrates the DAM as a post-processing unit after GCN within the normalizing flow framework. Simulations show that the proposed model is robust against noise and negative samples. Experimental results show that DA-Flow reaches competitive or better performance than the existing state-of-the-art (SOTA) methods in terms of the micro AUC metric with the fewest number of parameters. Moreover, we found that even without training, simply using random projection without dimensionality reduction on skeleton data enables substantial anomaly detection capabilities.</abstract><doi>10.48550/arxiv.2406.02976</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2406.02976
ispartof
issn
language eng
recordid cdi_arxiv_primary_2406_02976
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Computer Vision and Pattern Recognition
title DA-Flow: Dual Attention Normalizing Flow for Skeleton-based Video Anomaly Detection
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-29T18%3A06%3A17IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=DA-Flow:%20Dual%20Attention%20Normalizing%20Flow%20for%20Skeleton-based%20Video%20Anomaly%20Detection&rft.au=Wu,%20Ruituo&rft.date=2024-06-05&rft_id=info:doi/10.48550/arxiv.2406.02976&rft_dat=%3Carxiv_GOX%3E2406_02976%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true