一种基于SVM主动学习的卡通视频检测方法

高新波; 田春娜; 张娜

doi:10.3724/SP.J.1146.2005.01193

一种基于SVM主动学习的卡通视频检测方法

doi: 10.3724/SP.J.1146.2005.01193

基金项目:

新世纪优秀人才支持计划(NCET-04-0948)，教育部重点项目(10417 3)，国家自然科学基金(60202004)和教育部留学归国人员实验室资助课题

计量
- 文章访问数: 2842
- HTML全文浏览量: 93
- PDF下载量: 1824
- 被引次数: 0
出版历程
- 收稿日期: 2005-09-19
- 修回日期: 2006-11-22
- 刊出日期: 2007-06-19

A Cartoon Video Detection Method Based on Active SVM Learning

摘要

摘要: 通过分析卡通与非卡通视频在视觉上的差异，对视频片断提取了MPEG-7描述子等8组视觉特征来构造卡通视频的特征空间；并将主动相关反馈技术引入到支撑向量机(SVM)算法中，设计了一种基于主动学习的卡通视频检测分类方法。利用大量实际视频片断所做的测试实验结果表明，该文选取的特征对卡通和非卡通视频有较好的区分能力；且与单纯的SVM算法以及传统相关反馈和SVM算法结合的方法相比，该文算法在检测性能上有较大的优势。
- 视频检测;MPEG-7;支撑向量机;相关反馈;主动相关反馈
Abstract: Through analyzing visual differences between cartoon and other videos, 8 groups of typical features including MPEG-7 descriptors are extracted to construct the feature space of cartoon videos. Then, a content-based video classifier is designed by introducing the active relevance feedback technique into Support Vector Machine (SVM) for the cartoon video detection. Experimental results on a great many real video clips illustrate that the constructed feature space can represent the cartoon videos effectively. In addition, compared with the classifier based on SVM and that based on the traditional relevance feedback technique and SVM, the proposed classifier has a higher performance for cartoon video detection.

HTML全文

参考文献(1)

Fischer S, Lienhart R, and Effelsberg W. Automatic recognition of film genres. The 3rd ACM International Multimedia Conference and Exhibition. San Francisco, California, USA, Nov. 5-9, 1995, 1: 295-304.[2]Ianeva T I, de Vries A P, and Rohrig H. Detecting cartoons: A case study in automatic video-genre classification. Proc. IEEE International Conference on Multimedia and Expo. Baltimore, Maryland, Jul. 6-9, 2003, 1: 449-452.[3]Roach M, Mason J S, and Pawlewski M. Motion-based classification of cartoons. International Conference on Intelligent Multimedia, Video and Speech Processing, Hong Kong, May 2-4, 2001: 146-149.[4]Hyvarinen A. Fast and robust fixed-point algorithms for independent component analysis[J].IEEE Trans. on Neural Networks.1999, 10(3):626-634[5]田春娜, 高新波, 李洁. 基于嵌入式Bootstrap的主动学习示例选择方法. 计算机研究与发展, 2006, 43(10): 1706-1712. Tian C N, Gao X B, and Li J. An example selection method for active learning based on embedded Bootstrap algorithm. Journal of Computer Research and Development, 2006, 43(10): 1706-1712.[6]陈可佳, 姜远, 周志华. 基于主动相关反馈的图像检索方法. 模式识别与人工智能, 2005, 18(4): 480-485. Chen K J, Jiang Y, and Zhou Z H. An image retrieval method based on active relevance feedback. Pattern Recognition and Artificial Intelligence, 2005, 18(4): 480-485.[7]Zhang L, Lin F Z, and Zhang B. A CBIR method based on color-spatial feature. Proceedings of the IEEE Region 10 Conference on TENCON. Korea. Sept. 15-17, 1999, 1: 166-169.[8]Manjunath B S, Ohm J R, and Vasudevan V V, et al.. Color and texture descriptors[J].IEEE Trans. on Circuits and Systems for Video Technology.2001, 11(6):703-715[9]Ro Y M, Kim M, and Kang K, et al.. MPEG-7 homogeneous texture descriptor[J].ETRI Journal.2001, 23(2):41-51[10]Hasler D and Susstrunk S. Measuring colourfulness in natural images. Proc. SPIE/IST Human Vision and Electronic Imaging. Santa Clara, California, USA, Jan. 20-24, 2003: 87-95.[11]Burges C J C. A tutorial on support vector machines for pattern recognition[J].Knowledge Discovery and Data Mining.1998, 2(2):121-167[12]Gao X B and Tang X. Unsupervised video shot segmentation and model-free anchorperson detection for news video story parsing, IEEE Trans[J].on Circuits Systems for Video Technology.2002, 12(9):765-776

施引文献

资源附件(0)

访问统计