Video Sequence Boundary Detection Using Neural Gas Networks

Video sequence boundary detection is an important first step in the construction of efficient and user-friendly video archives. In this paper, we propose to employ growing neural gas (GNG) networks [7] to detect the shot boundaries, as the neural networks are capable of learning the characteristics of various shots andclustering them accordingly. We represent the image frames by 6-bit color-coded histograms.We make use of the chi-square distances between histograms of neighboring frames as the primary features to train the GNG and to detect the shot boundaries. Experimental results presentedin this paper demonstrate the reliable performance of our proposedapproac h on real video sequences.

[1]  Ralph M. Ford,et al.  Metrics for shot boundary detection in digital video sequences , 2000, Multimedia Systems.

[2]  Thomas S. Huang,et al.  Browsing and retrieving video content in a unified framework , 1998, 1998 IEEE Second Workshop on Multimedia Signal Processing (Cat. No.98EX175).

[3]  David Bordwell,et al.  Film Art: An Introduction , 1979 .

[4]  Bernd Fritzke,et al.  Growing cell structures--A self-organizing network for unsupervised and supervised learning , 1994, Neural Networks.

[5]  Akio Nagasaka,et al.  Automatic Video Indexing and Full-Video Search for Object Appearances , 1991, VDB.

[6]  T. Martínez,et al.  Competitive Hebbian Learning Rule Forms Perfectly Topology Preserving Maps , 1993 .

[7]  Bernd Fritzke,et al.  A Growing Neural Gas Network Learns Topologies , 1994, NIPS.

[8]  Riccardo Leonardi,et al.  Scene break detection: a comparison , 1998, Proceedings Eighth International Workshop on Research Issues in Data Engineering. Continuous-Media Databases and Applications.

[9]  Stefano Spaccapietra,et al.  Visual Database Systems 3 , 1995, IFIP — The International Federation for Information Processing.