Fast Hierarchical Multimodal Structuring of Time Slots

TV stream structuring consists primarily in breaking down the stream into advertisings, telecasts and time slots. These time slots and telecasts need then to be structured. This paper focuses on the structuring of one of the dayparts: the morning drive. The challenge resides in these morning drives' components heterogeneity under a same theme and branding: music, weather, news, sets, advertisings and reports are mixed. Classical structuring methods concern videos with homogeneous components, like sport games, and need few features. We propose here a solution to time slots structuring by modeling morning drives' structure in order to decrease the number of needed detections and to decrease the number of false alarms. Good results are obtained since 97% of the segments are correctly retrieved and since the accuracy of the retrieved structure is approximately 5 seconds.

[1]  Marcel Worring,et al.  Multimodal Video Indexing : A Review of the State-ofthe-art , 2001 .

[2]  Nuno Vasconcelos,et al.  Statistical models of video structure for content analysis and characterization , 2000, IEEE Trans. Image Process..

[3]  Patrick Gros,et al.  Multimodal Segmental-Based Modeling of Tennis Video Broadcasts , 2005, 2005 IEEE International Conference on Multimedia and Expo.

[4]  Jean Carrive,et al.  Improving Program Guides for Reducing TV Stream Structuring Problem to a Simple Alignment Problem , 2006, 2006 International Conference on Computational Inteligence for Modelling Control and Automation and International Conference on Intelligent Agents Web Technologies and International Commerce (CIMCA'06).

[5]  Patrick Gros,et al.  Hierarchical structure analysis of sport videos using HMMS , 2003, Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429).

[6]  Jean Carrive,et al.  Modeling Television Schedules for Television Stream Structuring , 2007, MMM.

[7]  Thomas S. Huang,et al.  Constructing table-of-content for videos , 1999, Multimedia Systems.

[8]  Shih-Fu Chang,et al.  Structure analysis of soccer video with hidden Markov models , 2002, 2002 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[9]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[10]  J. Rissanen.  A UNIVERSAL PRIOR FOR INTEGERS AND ESTIMATION BY MINIMUM DESCRIPTION LENGTH , 1983 .

[11]  Frédéric Bimbot,et al.  Experiments on speech tracking in audio documents using Gaussian mixture modeling , 2001, 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221).

[12]  A. Kosmala,et al.  A New Approach To Content-Based Video Indexing Using Hidden Markov Models , 1997 .

[13]  Jean Carrive,et al.  Television Stream Structuring with Program Guides , 2006, Eighth IEEE International Symposium on Multimedia (ISM'06).

[14]  Julien Pinquier,et al.  Audio indexing: primary components retrieval , 2006, Multimedia Tools and Applications.