Title | ||
---|---|---|
Decomposing the video editing structure of a talk-show using nonnegative matrix factorization |
Abstract | ||
---|---|---|
We introduce a novel video structuring scheme that exploits nonnegative matrix factorization (NMF) on count data (in a bag of features representation of the visual stream) to jointly discover latent structuring patterns and their activations in time. Our NMF variant employs the Kullback-Leibler divergence as a cost function and imposes a temporal smoothness constraint to the activations. It is solved by a majorization-minimization technique. Our method is shown to be successful for decomposing the high-level editing structure of talk-shows. It is evaluated using a challenging database of TV political-debate programs, and found to clearly outperform a reference HMM method. |
Year | DOI | Venue |
---|---|---|
2012 | 10.1109/ICIP.2012.6467557 | Image Processing |
Keywords | Field | DocType |
image representation,matrix decomposition,minimisation,video signal processing,Kullback-Leibler divergence,NMF variant,TV political-debate programs,cost function,count data,features representation,latent structuring patterns,majorization-minimization technique,nonnegative matrix factorization,reference HMM method,talk-show,temporal smoothness constraint,video editing structure decomposition,visual stream,Video structuring,bag of features,indexing,machine learning,matrix factorization,unsupervised classification | Pattern recognition,Computer science,Matrix decomposition,Video editing,Minimisation (psychology),Non-negative matrix factorization,Count data,Artificial intelligence,Smoothness,Structuring,Hidden Markov model | Conference |
ISSN | ISBN | Citations |
1522-4880 E-ISBN : 978-1-4673-2532-5 | 978-1-4673-2532-5 | 1 |
PageRank | References | Authors |
0.36 | 6 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Slim Essid | 1 | 212 | 32.00 |
Cédric Févotte | 2 | 2380 | 149.37 |