{"quality_controlled":0,"title":"Topic models for semantic video compression","status":"public","publist_id":"2705","date_created":"2018-12-11T12:04:34Z","year":"2010","publisher":"ACM","publication_status":"published","citation":{"short":"J. Wanke, A. Ulges, C. Lampert, T. Breuel, in:, ACM, 2010, pp. 275–284.","ama":"Wanke J, Ulges A, Lampert C, Breuel T. Topic models for semantic video compression. In: ACM; 2010:275-284. doi:<a href=\"https://doi.org/10.1145/1743384.1743433\">10.1145/1743384.1743433</a>","ista":"Wanke J, Ulges A, Lampert C, Breuel T. 2010. Topic models for semantic video compression. MIR: Multimedia Information Retrieval, 275–284.","chicago":"Wanke, Jörn, Adrian Ulges, Christoph Lampert, and Thomas Breuel. “Topic Models for Semantic Video Compression,” 275–84. ACM, 2010. <a href=\"https://doi.org/10.1145/1743384.1743433\">https://doi.org/10.1145/1743384.1743433</a>.","apa":"Wanke, J., Ulges, A., Lampert, C., &#38; Breuel, T. (2010). Topic models for semantic video compression (pp. 275–284). Presented at the MIR: Multimedia Information Retrieval, ACM. <a href=\"https://doi.org/10.1145/1743384.1743433\">https://doi.org/10.1145/1743384.1743433</a>","ieee":"J. Wanke, A. Ulges, C. Lampert, and T. Breuel, “Topic models for semantic video compression,” presented at the MIR: Multimedia Information Retrieval, 2010, pp. 275–284.","mla":"Wanke, Jörn, et al. <i>Topic Models for Semantic Video Compression</i>. ACM, 2010, pp. 275–84, doi:<a href=\"https://doi.org/10.1145/1743384.1743433\">10.1145/1743384.1743433</a>."},"date_published":"2010-03-31T00:00:00Z","extern":1,"conference":{"name":"MIR: Multimedia Information Retrieval"},"_id":"3676","abstract":[{"lang":"eng","text":"Most state-of-the-art systems for content-based video understanding tasks require video content to be represented as collections of many low-level descriptors, e.g. as histograms of the color, texture or motion in local image regions.\n\nIn order to preserve as much of the information contained in the original video as possible, these representations are typically high-dimensional, which conflicts with the aim for compact descriptors that would allow better efficiency and lower storage requirements.\nIn this paper, we address the problem of semantic com- pression of video, i.e. the reduction of low-level descriptors to a small number of dimensions while preserving most of the semantic information. For this, we adapt topic models – which have previously been used as compact representations of still images – to take into account the temporal structure of a video, as well as multi-modal components such as motion information.\n\nExperiments on a large-scale collection of YouTube videos show that we can achieve a compression ratio of 20 : 1 compared to ordinary histogram representations and at least 2 : 1 compared to other dimensionality reduction techniques without significant loss of prediction accuracy. Also, improvements are demonstrated for our video-specific extensions modeling temporal structure and multiple modalities."}],"main_file_link":[{"open_access":"0","url":"http://pub.ist.ac.at/~chl/papers/wanke-mir2010.pdf"}],"type":"conference","day":"31","author":[{"full_name":"Wanke,Jörn","first_name":"Jörn","last_name":"Wanke"},{"last_name":"Ulges","first_name":"Adrian","full_name":"Ulges, Adrian"},{"full_name":"Christoph Lampert","id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","last_name":"Lampert","first_name":"Christoph","orcid":"0000-0001-8622-7887"},{"full_name":"Breuel,Thomas M","last_name":"Breuel","first_name":"Thomas"}],"page":"275 - 284","doi":"10.1145/1743384.1743433","month":"03","date_updated":"2021-01-12T07:45:04Z"}