[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117984-en":3,"doc-seo-117984-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},117984,4810365810221,"Aurora","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Anomaly Detection in Compressed Video Streams Using Machine Learning - Technical disclosure commons article","Video outlier or anomaly detection currently uses computer vision and image processing on decompressed video streams, which is computationally expensive. This disclosure enables anomaly identification directly from compressed video bitstreams using a pre-trained machine learning model, avoiding full decompression. Extracted compressed-domain features feed the model to produce a strong anomaly signal and optionally trigger alerts. When anomalies are indicated, only the relevant sequences are decompressed for deeper analysis. Sequences with ground truth can further support continuous unsupervised training to improve future detection without new labeled data or separate training cycles.","Technical Disclosure Commons  \nDefensive Publications Series  \nOctober 2023  \nAnomaly Detection in Compressed Video Streams Using Machine Learning  \nYuriy Aleksandrovich Romanenko  \nFollow this and additional works at: [https://www.tdcommons.org/dpubs_series](https://www.tdcommons.org/dpubs_series)  \nRecommended Citation  \nRomanenko, Yuriy Aleksandrovich, \"Anomaly Detection in Compressed Video Streams Using Machine Learning\", Technical Disclosure Commons,(October 23, 2023)  \n[https://www.tdcommons.org/dpubs_series/6340](https://www.tdcommons.org/dpubs_series/6340)  \nThis work is licensed under a Creative Commons Attribution 4.0 License.  \nThis Article is brought to you for free and open access by Technical Disclosure Commons. It has been accepted for inclusion in Defensive Publications Series by an authorized administrator of Technical Disclosure Commons.  \nAnomaly Detection in Compressed Video Streams Using Machine Learning  \nABSTRACT  \nVideo outlier or anomaly detection is currently performed using computer vision and  \nimage processing techniques such as filtering, motion estimation, etc. that require  \ndecompressed video streams. Applying machine learning techniques directly to a video stream  \nis computationally expensive. This disclosure describes the use of a pre-trained machine  \nlearning model that can analyze compressed video bitstreams directly without decompression to identify likely anomalies in the video. Such direct analysis of compressed video streams can be  \nperformed at significantly lower computational cost while still yielding a strong anomaly  \ndetection signal. Sequences in which anomalies are detected can be decompressed and further  \nanalyzed using conventional techniques of video analysis. Sequences over longer time spans  \n(and associated groundtruth information about whether an anomaly exists in the sequence) can be used for continuous unsupervised training of the same model. This can improve detection of  \nanomalies in subsequent sequences to enable better anomaly detection, without requiring  \ncollection of additional training data or separate training cycles.  \nKEYWORDS  \n● Anomaly detection  \n● Video anomaly  \n● Compressed bitstream  \n● Surveillance camera  \n● Security camera  \n● Unsupervised learning  \n● Anomaly detection  \nPublished by Technical Disclosure Commons, 2023 2  \nBACKGROUND  \nVideo outlier or anomaly detection is important in many applications. For example,  \nanalysis of video from a surveillance camera or other video recording devices can reveal  \nintrusion or other actions that are unusual (deviate from normal patterns) . Currently, video  \nanomaly detection relies on computer vision and image processing techniques such as filtering,  \nmotion estimation, etc. applied to decompressed video streams. Applying machine learning  \ntechniques directly to a video stream is computationally expensive.  \nDESCRIPTION  \nVideo streams produced by a camera (e.g., surveillance camera or other cameras) maybe transmitted as a video stream to a digital video recorder (DVR) or network video recorder (NVR) for storage. Such streams are usually compressed using video codecs that utilize imageto-image prediction, motion estimation, etc. This disclosure describes machine learning  \ntechniques that can analyze the compressed bitstreams directly without decompression to  \nidentify likely anomalies in the video. Such direct analysis of compressed video streams can be  \nperformed at significantly lower computational cost while still yielding a strong anomaly  \ndetection signal.  \n[https://www.tdcommons.org/dpubs_series/6340](https://www.tdcommons.org/dpubs_series/6340) 3  \nFig. 1: Anomaly detection in compressed video stream using machine learning  \nFig. 1 illustrates a method to identify anomalies in compressed video streams using a machine learning model. A compressed video bitstream is parsed but not decompressed (102) . Sequences of compressed picture data are extracted from the parsed stream. These can ","cbCairad57if9kfA","https://ap.wps.com/l/cbCairad57if9kfA","pdf",157019,1,6,"English","en",105,"# Abstract\n# Background\n# Description\n## Compressed-domain anomaly detection workflow\n## Optional alerting and selective decompression\n## Continuous unsupervised training","[{\"question\":\"Why is direct machine learning on video streams challenging with existing approaches?\",\"answer\":\"Conventional anomaly detection often relies on computer vision and image processing that require decompressed video streams, which makes direct machine learning computationally expensive.\"},{\"question\":\"How does the disclosed method detect anomalies without decompression?\",\"answer\":\"The method parses a compressed video bitstream, extracts sequences and compressed-domain metadata/features, and feeds them into a pre-trained machine learning model to output whether an anomaly likely exists.\"},{\"question\":\"What happens when the model indicates an anomaly?\",\"answer\":\"An alert can be raised. Additionally, only the sequences flagged by the compressed-bitstream analysis are decompressed for in-depth conventional video analysis.\"}]","Anomaly Detection in Compressed Video Streams Using Machine Learning - Technical disclosure commons article | PDF",1785680641,15,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"anomaly-detection-in-compressed-video-streams-using-machine-learning-technical-disclosure-commons-article","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/anomaly-detection-in-compressed-video-streams-using-machine-learning-technical-disclosure-commons-article/117984/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why is direct machine learning on video streams challenging with existing approaches?","Question",{"text":75,"@type":76},"Conventional anomaly detection often relies on computer vision and image processing that require decompressed video streams, which makes direct machine learning computationally expensive.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the disclosed method detect anomalies without decompression?",{"text":80,"@type":76},"The method parses a compressed video bitstream, extracts sequences and compressed-domain metadata/features, and feeds them into a pre-trained machine learning model to output whether an anomaly likely exists.",{"name":82,"@type":73,"acceptedAnswer":83},"What happens when the model indicates an anomaly?",{"text":84,"@type":76},"An alert can be raised. Additionally, only the sequences flagged by the compressed-bitstream analysis are decompressed for in-depth conventional video analysis.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]