Filters








267 Hits in 4.1 sec

Fine-grained hidden markov modeling for broadcast-news story segmentation

Warren Greiff, Alex Morgan, Randall Fish, Marc Richards, Amlan Kundu
2001 Proceedings of the first international conference on Human language technology research - HLT '01   unpublished
We present the design and development of a Hidden Markov Model for the division of news broadcasts into story segments.  ...  Model topology, and the textual features used, are discussed, together with the non-parametric estimation techniques that were employed for obtaining estimates for both transition and observation probabilities  ...  In this paper we discuss an approach to segmentation based on the use of a fine-grained Hidden Markov Model [Rabiner, `89 ] to model the generation of the words produced during a news program.  ... 
doi:10.3115/1072133.1072181 fatcat:dn3445n235a7xnxlj3qnlsycbe

The CIMWOS Multimedia Indexing System [chapter]

Harris Papageorgiou, Athanassios Protopapas
2004 Lecture Notes in Computer Science  
The database includes sports, broadcast news and documentaries in three languages.  ...  The CIMWOS Retrieval Engine is based on a weighted boolean model with intelligent indexing components.  ...  Story Detection (SD) and Topic Classification (TC) The basis of the Story Detection (SD) and Topic Classification (TC) modules is a generative mixture-based Hidden Markov Model (HMM).  ... 
doi:10.1007/978-3-540-27814-6_72 fatcat:6u2oul5ucbbwraqqek7hnzj7ey

Topic Segmentation [chapter]

Matthew Purver
2011 Spoken Language Understanding  
This chapter discusses the task of topic segmentation: automatically dividing single long recordings or transcripts into shorter, topically coherent segments.  ...  The most commonly used generative model in ASR is the hidden Markov model (HMM), and Figure 1 .7(a) shows a HMM can be used as a simple topic model.  ...  However, producing a fine-grained segmentation turns out to be an extremely difficult task.  ... 
doi:10.1002/9781119992691.ch11 fatcat:lvloaozyaraupna4fmy4kfhd3i

Prosody-Based Automatic Segmentation of Speech into Sentences and Topics [article]

E. Shriberg and A. Stolcke and D. Hakkani-Tur and G. Tur
2000 arXiv   pre-print
Using decision tree and hidden Markov modeling techniques, we combine prosodic cues with word-based approaches, and evaluate performance on two speech corpora, Broadcast News and Switchboard.  ...  For example, pause and pitch features are highly informative for segmenting news speech, whereas pause, duration and word-based cues dominate for natural conversation.  ...  Acknowledgements We thank Kemal Sönmez for providing the model for F0 stylization used in this work; Rebecca Bates, Mari Ostendorf, Ze'ev Rivlin, Ananth Sankar, and Ke-mal Sönmez for invaluable assistance  ... 
arXiv:cs/0006036v1 fatcat:zmnqcflvmfgrbj6gdcmj3j2cim

Forecasting Events Using an Augmented Hidden Conditional Random Field [chapter]

Xinyu Wei, Patrick Lucey, Stephen Vidas, Stuart Morgan, Sridha Sridharan
2015 Lecture Notes in Computer Science  
However, as the prediction is based solely on the hidden layer, fine-grained local information is not incorporated which reduces its predictive capability.  ...  For forecasting complex events, higher-order models such as Hidden Conditional Random Field (HCRF) have been used to good effect as capture the long-term, high-level semantics of the signal.  ...  This modification allows our model to not only capture a coarse summarization of what has happen so far through the hidden layer but also include fine-grained information of the current situation via the  ... 
doi:10.1007/978-3-319-16817-3_37 fatcat:jdzyr3rijbal7l4xyofl2eoyxi

Topic segmentation of TV-streams by watershed transform and vectorization

Vincent Claveau, Sébastien Lefèvre
2015 Computer Speech and Language  
A fine-grained segmentation of Radio or TV broadcasts is an essential step for most multimedia processing tasks. Applying segmentation algorithms to the speech transcripts seems straightforward.  ...  This new topic segmentation technique is evaluated on two corpora of transcripts from French TV broadcasts on which it largely outperforms other existing approaches from the state-of-the-art.  ...  Bellot (LIA, France) for making their datasets available for our experiments, and for the seminal discussion about TV stream segmentation.  ... 
doi:10.1016/j.csl.2014.04.006 fatcat:s6ewbijwbjdzte4lsi7ia6bh74

Analysis of implemented part of speech tagger approaches: The case of Ethiopian languages

Wubetu Barud Demilie, Department of Information Technology, Wachemo University, Hossana, Ethiopia, P.O. Box 667
2020 Indian Journal of Science and Technology  
For example, it is possible to see the English version of the Hidden Markov Model-based POS tagger using Stanford tagger/parser (5) (6) . Example: "We can can the can."  ...  It likewise settles a few ambiguities that are not tended to by the syntactic parser's language model (3) (4) .  ...  Acknowledgment The author wants to acknowledge all researchers of the area who have been contributed a lot regardless of POS tagging research work for Ethiopian languages.  ... 
doi:10.17485/ijst/v13i48.1876 fatcat:hpsvixt5tnamplkuk2niyfn2lu

Content-based inference of hierarchical structural grammar for recurrent TV programs using multiple sequence alignment

Bingqing Qu, Felicien Vallet, Jean Carrive, Guillaume Gravier
2014 2014 IEEE International Conference on Multimedia and Expo (ICME)  
Hidden Markov models have been widely used in the first case [6] , [7] , while event detection has been applied to both cases.  ...  IEEE, 1998. [5] Xinbo Gao and Xiaoou Tang, "Unsupervised video-shot segmentation and model-free anchorperson detection for news video story parsing," IEEE Transactions on Circuits and Systems for Video  ... 
doi:10.1109/icme.2014.6890295 dblp:conf/icmcs/QuVCG14 fatcat:oascgn4csvh3fciyvm3eatziha

Integrating Prosodic and Lexical Cues for Automatic Topic Segmentation

Gökhan Tür, Dilek Hakkani-Tür, Andreas Stolcke, Elizabeth Shriberg
2001 Computational Linguistics  
We propose two methods for combining lexical and prosodic information using hidden Markov models and decision trees.  ...  We evaluate our approach on the Broadcast News corpus, using the DARPA-TDT evaluation metrics. Results show that the prosodic model alone is competitive with word-based segmentation methods.  ...  Acknowledgments We thank Becky Bates, Madelaine PlauchG Ze'ev Rivlin, Ananth Sankar, and Kemal S6nmez for invaluable assistance in preparing the data for this study.  ... 
doi:10.1162/089120101300346796 fatcat:cnz4bvenjbgq3jwzts2zwe5224

Integrating Prosodic and Lexical Cues for Automatic Topic Segmentation [article]

G. Tur, D. Hakkani-Tur, A. Stolcke, E. Shriberg
2001 arXiv   pre-print
We propose two methods for combining lexical and prosodic information using hidden Markov models and decision trees.  ...  We evaluate our approach on the Broadcast News corpus, using the DARPA-TDT evaluation metrics. Results show that the prosodic model alone is competitive with word-based segmentation methods.  ...  Acknowledgments We thank Becky Bates, Madelaine Plauché, Ze'ev Rivlin, Ananth Sankar, and Kemal Sönmez for invaluable assistance in preparing the data for this study.  ... 
arXiv:cs/0105037v1 fatcat:ybpzxrbztfc3jk55bi77vkocme

Using acoustic structure in a hand-held audio playback device

C. Schmandt, D. Roy
1996 IBM Systems Journal  
News-Comm is a hand-held digital audio playback device, designed for portable listening, selecting, and scanning of radio newscasts and other audio "programs."  ...  NewsComm is oriented toward portable listening, such as during commute time, and uses a model of occasional network connectivity to an audio server that selects recordings for each user and applies signal  ...  This work was supported by the MIT News in the Future consortium. **Trademark or registered trademark of Sun Microsystems, Inc. or Intel Corp.  ... 
doi:10.1147/sj.353.0453 fatcat:np7y7zyfvbcrnnp5ojlxu3wqfi

Toward a Better Understanding of News User Journeys: A Markov Chain Approach

Susan Vermeer, Damian Trilling
2020 Journalism Studies  
for their website.  ...  To move forward, we demonstrate a computational approach to understand the news journeys Web users take to find the news they want to read. We propose the use of Markov chains.  ...  They tackle the problem of overreporting in surveys on news use (see Prior 2009) , and allow for a much more fine-grained level of analysis.  ... 
doi:10.1080/1461670x.2020.1722958 fatcat:5gutfeui5jhzdgadhhv4fgycyy

Content-based unsupervised segmentation of recurrent TV programs using grammatical inference

Bingqing Qu, Félicien Vallet, Jean Carrive, Guillaume Gravier
2017 Multimedia tools and applications  
A model of the structure is derived from the structural grammars and used to predict the structure of new episodes. Comparative evaluation on two gram-Bingqing Qu  ...  TV program segmentation raised as a major topic in the last decade for the task of high quality indexing of multimedia content.  ...  turns the problem to new research directions, e.g., a mixture model generalized by a hidden Markov model.  ... 
doi:10.1007/s11042-017-4816-5 fatcat:o5ogwy3bufdildpttwepkewwxa

Using MPEG Standards for Content-Based Indexing of Broadcast Television, Web, and Enterprise Content [chapter]

David Gibbon, Zhu Liu, Andrea Basso, Behzad Shahraray
2010 The Handbook of MPEG Applications  
Its use in DVDs, Direct Broadcast Satellite (DBS), Cable, Digital Video Broadcast (DVB) systems as well as the recent U.S. transition to digital TV (ATSC) ensures that MPEG-2 will remain in use for many  ...  Widely adopted standards are essential for these systems and devices to function seamlessly, not only at the video codec and transport level, but also with respect to metadata for program descriptions  ...  Relying on a set of 14 audio features extracted for each audio clip, a hidden Markov model (HMM) classifier achieves an accuracy of 84.5%.  ... 
doi:10.1002/9780470974582.ch14 fatcat:kp47vsowi5cvhg6mgw4s4uhte4

Joint modality fusion and temporal context exploitation for semantic video analysis

Georgios Th Papadopoulos, Vasileios Mezaris, Ioannis Kompatsiaris, Michael G. Strintzis
2011 EURASIP Journal on Advances in Signal Processing  
Then, Hidden Markov Models (HMMs) are employed for performing an initial association of each shot with the semantic classes that are of interest separately for each modality.  ...  Overall, the examined video sequence is initially segmented into shots and for every resulting shot appropriate color, motion and audio features are extracted.  ...  distribution information to Hidden Markov Models (HMMs).  ... 
doi:10.1186/1687-6180-2011-89 fatcat:xhvlcefqpzaprdkgb72l2hxlam
« Previous Showing results 1 — 15 out of 267 results