Correlative multi-label video annotation Export

@inproceedings{Qi2007multi-label, abstract = {Automatically annotating concepts for video is a key to semantic-level video browsing, search and navigation. The research on this topic evolved through two paradigms. The first paradigm used binary classification to detect each individual concept in a concept set. It achieved only limited success, as it did not model the inherent correlation between concepts, e.g., urban and building. The second paradigm added a second step on top of the individual concept detectors to fuse multiple concepts. However, its performance varies because the errors incurred in the first detection step can propagate to the second fusion step and therefore degrade the overall performance. To address the above issues, we propose a third paradigm which simultaneously classifies concepts and models correlations between them in a single step by using a novel Correlative Multi-Label (CML) framework. We compare the performance between our proposed approach and the state-of-the-art approaches in the first and second paradigms on the widely used TRECVID data set. We report superior performance from the proposed approach.}, address = {New York, NY, USA}, author = {Qi, Guo J. and Hua, Xian S. and Rui, Yong and Tang, Jinhui and Mei, Tao and Zhang, Hong J.}, booktitle = {MULTIMEDIA '07: Proceedings of the 15th international conference on Multimedia}, citeulike-article-id = {2506113}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1291245}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1291233.1291245}, doi = {10.1145/1291233.1291245}, isbn = {978-1-59593-702-5}, keywords = {learning, multi-label}, location = {Augsburg, Germany}, pages = {17--26}, posted-at = {2008-03-11 06:44:03}, priority = {2}, publisher = {ACM}, title = {Correlative multi-label video annotation}, url = {http://dx.doi.org/10.1145/1291233.1291245}, year = {2007} }

TY - CONF ID - Qi2007multi-label L3 - citeulike-article-id:2506113 N2 - Automatically annotating concepts for video is a key to semantic-level video browsing, search and navigation. The research on this topic evolved through two paradigms. The first paradigm used binary classification to detect each individual concept in a concept set. It achieved only limited success, as it did not model the inherent correlation between concepts, e.g., urban and building. The second paradigm added a second step on top of the individual concept detectors to fuse multiple concepts. However, its performance varies because the errors incurred in the first detection step can propagate to the second fusion step and therefore degrade the overall performance. To address the above issues, we propose a third paradigm which simultaneously classifies concepts and models correlations between them in a single step by using a novel Correlative Multi-Label (CML) framework. We compare the performance between our proposed approach and the state-of-the-art approaches in the first and second paradigms on the widely used TRECVID data set. We report superior performance from the proposed approach. CY - Augsburg, Germany SN - 978-1-59593-702-5 AD - New York, NY, USA EP - 26 TI - Correlative multi-label video annotation PB - ACM T2 - MULTIMEDIA '07: Proceedings of the 15th international conference on Multimedia SP - 17 KW - learning KW - multi-label AU - Qi, Guo AU - Hua, Xian AU - Rui, Yong AU - Tang, Jinhui AU - Mei, Tao AU - Zhang, Hong PY - 2007/// UR - http://dx.doi.org/10.1145/1291233.1291245 ER -