{"title":"Real-Time Activity Search of Surveillance Video","authors":"Greg Castañón, Venkatesh Saligrama, André-Louis Caron, Pierre-Marc Jodoin","doi":"10.1109/AVSS.2012.58","DOIUrl":null,"url":null,"abstract":"We present a fast and flexible content-based retrieval method for surveillance video. Designing a video search robust to uncertain activity duration, high variability in object shapes and scene content is challenging. We propose a two-step approach to video search. First, local motion features are inserted into an inverted index using locality-sensitive hashing (LSH). Second, we utilize a novel optimization approach based on edit distance to minimize temporal distortion, limited obscuration and imperfect queries. This approach assembles the local features stored in the index into a video segment which matches the query video. Pre-processing of archival video is performed in real-time, and retrieval speed scales as a function of the number of matches rather than video length. We demonstrate the effectiveness of the approach for counting, motion pattern recognition and abandoned object applications using a pair of challenging video datasets.","PeriodicalId":275325,"journal":{"name":"2012 IEEE Ninth International Conference on Advanced Video and Signal-Based Surveillance","volume":"26 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2012-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2012 IEEE Ninth International Conference on Advanced Video and Signal-Based Surveillance","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AVSS.2012.58","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 5
Abstract
We present a fast and flexible content-based retrieval method for surveillance video. Designing a video search robust to uncertain activity duration, high variability in object shapes and scene content is challenging. We propose a two-step approach to video search. First, local motion features are inserted into an inverted index using locality-sensitive hashing (LSH). Second, we utilize a novel optimization approach based on edit distance to minimize temporal distortion, limited obscuration and imperfect queries. This approach assembles the local features stored in the index into a video segment which matches the query video. Pre-processing of archival video is performed in real-time, and retrieval speed scales as a function of the number of matches rather than video length. We demonstrate the effectiveness of the approach for counting, motion pattern recognition and abandoned object applications using a pair of challenging video datasets.