Multimodal topic segmentation and classification of news video

S. Raaijmakers1, J. den Hartog2, J. Baan2
1Netherlands Organisation for Applied Scientific Research (TNO), Institute of Applied Physics & MediaMill Delft/Amsterdam, The Netherlands
2Netherlands Organisation for Applied Scientific Research (TNO), Institute of Applied Physics & MediaMill Delft/Amsterdam, Amsterdam, The Netherlands

Tóm tắt

In this paper we describe a model for multimodal topic segmentation and classification of Dutch news video. A focal topic of interest for the research reported here is the interaction between three different modalities (visual, auditory and textual information) in an integrated model for video analysis. We present a fully automated sequential feedback model for video analysis, where linguistic analysis is combined with visual information for the purposes of both segmentation and classification.

Từ khóa

#Information analysis #Feedback #Image segmentation #Physics #Machine assisted indexing #Hidden Markov models #Detectors #Switches #Histograms #Cameras

Tài liệu tham khảo

hanjalic, 2001, Dancers Deift Advanced News Retrieval System beeferman, 1999, Statistical Models for Text Segmentation Machine Learning, 34, 1 hearst, 1997, Texttiling: Segmenting text into multi-paragraph subtopic passages, Computational Linguistics, 23, 33 maybury, 1996, Segmentation, content extraction and visualization of broadcast news video using multistream analysis, ACM international conference on Multimedia jasinschi, 2001, Integrated Multimedia Processing for Topic Segmentation and Classification, 366 10.3115/1072133.1072181 10.1109/HICSS.1998.651708 pevzner, 2002, A Critique and Improvement of An Evaluation Metric for Text Segmentation 10.1007/s005300050138