The Hong Kong Polytechnic University at TRECVID 2007 BBC rushes summarization

ABSTRACT

In this paper, we propose the framework and methodology for BBC rushes summarization task of TRECVID 2007. We divide the entire task into three sub-tasks: shot segmentation; noise shot detection and removal; video summarization. We first segment the rushes to a set of video shots based on the accumulative difference between the consecutive frames. In the second sub-task, we classify the noise video shots to several categories and detect them individually using different feature space. A novel audio-based method has been proposed to detect the video shots including clapper board noise. Thirdly, we extract the most representative video clips from the useful video shots and remove the redundant ones by comparing the similarity. We analyze the performance of these methods, and demonstrate them in the BBC rushes summarization task.

REFERENCES

• 1: Paul Over , Alan F. Smeaton , Philip Kelly, The trecvid 2007 BBC rushes summarization evaluation pilot, Proceedings of the international workshop on TRECVID video summarization, p.1-15, September 28-28, 2007, Augsburg, Bavaria, Germany [doi>10.1145/1290031.1290032]
• 2: Kikukawa, T., and Kawafuchi, S. Development of an automatic summary editing system for the audio visual resources. Transactions of the Institute of Electronics, Information and Communication Engineers, 1992: J75-A (2).
• 3: HongJiang Zhang , Atreyi Kankanhalli , Stephen W. Smoliar, Automatic partitioning of full-motion video, Multimedia Systems, v.1 n.1, p.10-28, June 1993 [doi>10.1007/BF01210504]
• 4: Ma, Y.F., Sheng, J., Chen, Y., and Zhang, H.J. MSR-Asia at TREC-10 Video Track: Shot Boundary Detection Task. Video TREC. 2001.
• 5: Pfeiffer, S., Lienhart, R., Fischer, S., and Effelsberg, W. Abstracting Digital Movies Automatically. Journal of Visual Communication and Image Representation, 7, 4 (Dec.) pp: 345--353. 1996.
• 6: Yan, J. X, Dou, W. B., and Dong, Z. W. Time-Domain Detection Method of Transient Signals in Audio Coding. Journal of Electronics& Information Technology, vol. 28, no. 2 (Feb. 2006), (in Chinese)

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.10 Vision and Scene Understanding
            Subjects: Video analysis

General Terms:
Algorithms, Experimentation

Keywords:
BBC rushes summarization, audio transient, clapper board noise detection, video rushes editing