Best practices for learning video concept detectors from social media examples
Best practices for learning video concept detectors from social media examples
复制标题
从社交媒体示例中学习视频概念检测器的最佳实践
DOI：
10.1007/s11042-014-2056-5
复制
发表时间：
2014-05
影响因子：
3.6
通讯作者：
Cees Snoek
中科院分区：
文献类型：
--
作者：
Svetlana Kordumova;Xirong Li;Cees Snoek
从Flickr图像和YouTube视频等社交媒体源学习视频概念检测器，有可能解决视频搜索的各种概念查询。虽然许多人已经认识到了这一潜力，这一主题的进展也令人印象深刻，但我们认为，对于如何从社交媒体例子中学习有效的视频概念检测器来说，关键问题仍然悬而未决。作为回答这些问题的初步尝试，我们使用一个视频搜索引擎进行了一项实验研究，该引擎能够从社交媒体例子中学习概念检测器，无论是社交标记的视频还是社交标记的图像。在视频搜索引擎中，我们研究了三种正例选择策略、三种反例选择策略和三种学习策略。性能是根据具有挑战性的TRECVID 2012基准进行评估的，该基准由600小时的互联网视频组成。从实验中我们得到了四个最佳实践：(1)标记的图像比标记的视频是更好的学习视频概念的来源；(2)选择标记相关的正训练样本总是有益的；(3)选择相关的负面样本是有利的，对于视频和图像来源应该区别对待；(4)在学习之前用选择的相关训练数据学习概念检测器比在学习过程中结合相关性要好。我们视频搜索引擎中的最佳实践在TRECVID 2013基准中提供了最先进的性能，无需手动提供注释即可进行概念检测。
Learning video concept detectors from social media sources, such as Flickr images and YouTube videos, has the potential to address a wide variety of concept queries for video search. While the potential has been recognized by many, and progress on the topic has been impressive, we argue that key questions crucial to knowhow to learn effective video concept detectors from social media examples?remain open. As an initial attempt to answer these questions, we conduct an experimental study using a video search engine which is capable of learning concept detectors from social media examples, be it socially tagged videos or socially tagged images. Within the video search engine we investigate three strategies for positive example selection, three negative example selection strategies and three learning strategies. The performance is evaluated on the challenging TRECVID 2012 benchmark consisting of 600 h of Internet video. From the experiments we derive four best practices: (1) tagged images are a better source for learning video concepts than tagged videos, (2) selecting tag relevant positive training examples is always beneficial, (3) selecting relevant negative examples is advantageous and should be treated differently for video and image sources, and (4) learning concept detectors with selected relevant training data before learning is better then incorporating the relevance during the learning process. The best practices within our video search engine lead to state-of-the-art performance in the TRECVID 2013 benchmark for concept detection without manually provided annotations.
登录
查看更多内容
DOI：
10.1109/icme.2007.4284698
发表时间：
2007-07
期刊：
2007 IEEE International Conference on Multimedia and Expo
影响因子：
--
作者：
Mingjing Li
通讯作者：
Mingjing Li
DOI：
10.1109/cvpr.2008.4587630
发表时间：
2008-06
期刊：
2008 IEEE Conference on Computer Vision and Pattern Recognition
影响因子：
--
作者：
Subhransu Maji;A. Berg;Jitendra Malik
通讯作者：
Subhransu Maji;A. Berg;Jitendra Malik
DOI：
10.1145/1367497.1367542
发表时间：
2008-04
期刊：
--
影响因子：
--
作者：
Börkur Sigurbjörnsson;R. V. Zwol
通讯作者：
Börkur Sigurbjörnsson;R. V. Zwol
DOI：
10.1023/b:visi.0000029664.99615.94
发表时间：
2004-11-01
影响因子：
19.5
作者：
Lowe, DG
通讯作者：
Lowe, DG
DOI：
10.1145/1526709.1526757
发表时间：
2009-04
期刊：
--
影响因子：
--
作者：
Dong Liu;Xian-Sheng Hua;Linjun Yang;Meng Wang;HongJiang Zhang
通讯作者：
Dong Liu;Xian-Sheng Hua;Linjun Yang;Meng Wang;HongJiang Zhang