{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/a-correlation-based-feature-representation","title":"A Correlation Based Feature Representation for First-Person Activity Recognition","arxiv_id":"1711.05523","date":"2017-11-15","proceeding":null,"authors":["Reza Kahani","Alireza Talebpour","Ahmad Mahmoudi-Aznaveh"],"abstract":"In this paper, a simple yet efficient activity recognition method for\nfirst-person video is introduced. The proposed method is appropriate for\nrepresentation of high-dimensional features such as those extracted from\nconvolutional neural networks (CNNs). The per-frame (per-segment) extracted\nfeatures are considered as a set of time series, and inter and intra-time\nseries relations are employed to represent the video descriptors. To find the\ninter-time relations, the series are grouped and the linear correlation between\neach pair of groups is calculated. The relations between them can represent the\nscene dynamics and local motions. The introduced grouping strategy helps to\nconsiderably reduce the computational cost. Furthermore, we split the series in\ntemporal direction in order to preserve long term motions and better focus on\neach local time window. In order to extract the cyclic motion patterns, which\ncan be considered as primary components of various activities, intra-time\nseries correlations are exploited. The representation method results in highly\ndiscriminative features which can be linearly classified. The experiments\nconfirm that our method outperforms the state-of-the-art methods on recognizing\nfirst-person activities on the two challenging first-person datasets.","url_abs":"http://arxiv.org/abs/1711.05523v2","url_pdf":"http://arxiv.org/pdf/1711.05523v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"a-correlation-based-feature-representation","repo_url":"https://github.com/rkahani/FirstPersonActivityRecognition","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"activity-recognition","task_name":"Activity Recognition"},{"task_slug":"egocentric-activity-recognition","task_name":"Egocentric Activity Recognition"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}