{"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/online-object-tracking-learning-and-parsing","title":"Online Object Tracking, Learning and Parsing with And-Or Graphs","arxiv_id":"1509.08067","date":"2015-09-27","proceeding":"CVPR 2014 6","authors":["Tianfu Wu","Yang Lu","Song-Chun Zhu"],"abstract":"This paper presents a method, called AOGTracker, for simultaneously tracking,\nlearning and parsing (TLP) of unknown objects in video sequences with a\nhierarchical and compositional And-Or graph (AOG) representation. %The AOG\ncaptures both structural and appearance variations of a target object in a\nprincipled way. The TLP method is formulated in the Bayesian framework with a\nspatial and a temporal dynamic programming (DP) algorithms inferring object\nbounding boxes on-the-fly. During online learning, the AOG is discriminatively\nlearned using latent SVM to account for appearance (e.g., lighting and partial\nocclusion) and structural (e.g., different poses and viewpoints) variations of\na tracked object, as well as distractors (e.g., similar objects) in background.\nThree key issues in online inference and learning are addressed: (i)\nmaintaining purity of positive and negative examples collected online, (ii)\ncontroling model complexity in latent structure learning, and (iii) identifying\ncritical moments to re-learn the structure of AOG based on its intrackability.\nThe intrackability measures uncertainty of an AOG based on its score maps in a\nframe. In experiments, our AOGTracker is tested on two popular tracking\nbenchmarks with the same parameter setting: the TB-100/50/CVPR2013 benchmarks,\nand the VOT benchmarks --- VOT 2013, 2014, 2015 and TIR2015 (thermal imagery\ntracking). In the former, our AOGTracker outperforms state-of-the-art tracking\nalgorithms including two trackers based on deep convolutional network. In the\nlatter, our AOGTracker outperforms all other trackers in VOT2013 and is\ncomparable to the state-of-the-art methods in VOT2014, 2015 and TIR2015.","url_abs":"http://arxiv.org/abs/1509.08067v6","url_pdf":"http://arxiv.org/pdf/1509.08067v6.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":"online-object-tracking-learning-and-parsing","repo_url":"https://github.com/tfwu/RGM-AOGTracker","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"object-tracking","task_name":"Object Tracking"}],"methods":[{"method_slug":"svm","method_name":"SVM"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1509.08067","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}