Phased Scene Change Detection in Ubiquitous Environments
In a ubiquitous environment, video transfer is very important. In particular, transferring of selected video clips obtained through scene change detection is more important than transferring an entire video. In this paper, inter-frame difference values are first computed through combining the χ 2 histogram with the color histogram, as well as normalization. Next, key frames for a cluster are determined through distance clustering and K-mean clustering. Lastly, key frames for a group are determined through a likelihood ratio. According to our experiments, the proposed method surpassed other methods in its ability to detect scene changes due to the use of three steps: difference value calculation, clustering, and key frame extraction.
KeywordsClustering Likelihood Ratio Key Frame K-mean Clustering
Unable to display preview. Download preview PDF.
- 5.Yeo, B.-L.: Efficiency processing of compressed image and video. Technical report, PhD thesis, Princeton University (1996)Google Scholar
- 6.Zhang, H., Low, C.Y., Gong, Y., Smoliar, S.: Video Parsing Using Compressed Data. In: Proc. SPIE Conf. Image and Video Processing II, vol. 2182, pp. 142–149 (1994)Google Scholar
- 7.Zhong, D., Chang, S.F.: Video Object Model and Segmentation for Content-Based Video Indexing. In: IEEE International Conference on Circuits and Systems (June 1997)Google Scholar
- 8.Yeung, M.M., Yeo, B.-L., Wolf, W., Liu, B.: Video Browsing Using Clustering and Scene Transition Compressed Sequences. IS&T SPIE, Multimedia Computing and Networking (1995)Google Scholar
- 10.Shin, S.Y., Pyo, S.B.: Video Browsing Using An Efficient Scene Change Detection in Telematics. Journal of KSCI 11(4), 147–154 (2006)Google Scholar
- 11.Kim, Y.L., Rhee, Y.W.: Scene Change Detection Using Local χ 2-Test. Journal of KSCI 11(3), 193–202 (2006)Google Scholar
- 12.Sarah, V.P.: Video Segmentation and indexing using Motion Estimation. Doctorial Dissertation, University of Bristol (2004)Google Scholar