Background Subtraction with Dirichlet Process Mixture Models. Haines, T. & Xiang, T. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(4):670--683, April, 2014. 00005doi abstract bibtex Video analysis often begins with background subtraction. This problem is often approached in two steps-a background model followed by a regularisation scheme. A model of the background allows it to be distinguished on a per-pixel basis from the foreground, whilst the regularisation combines information from adjacent pixels. We present a new method based on Dirichlet process Gaussian mixture models, which are used to estimate per-pixel background distributions. It is followed by probabilistic regularisation. Using a non-parametric Bayesian method allows per-pixel mode counts to be automatically inferred, avoiding over-/under- fitting. We also develop novel model learning algorithms for continuous update of the model in a principled fashion as the scene changes. These key advantages enable us to outperform the state-of-the-art alternatives on four benchmarks.
@article{ haines_background_2014,
title = {Background {Subtraction} with {Dirichlet} {Process} {Mixture} {Models}},
volume = {36},
issn = {0162-8828},
doi = {10.1109/TPAMI.2013.239},
abstract = {Video analysis often begins with background subtraction. This problem is often approached in two steps-a background model followed by a regularisation scheme. A model of the background allows it to be distinguished on a per-pixel basis from the foreground, whilst the regularisation combines information from adjacent pixels. We present a new method based on Dirichlet process Gaussian mixture models, which are used to estimate per-pixel background distributions. It is followed by probabilistic regularisation. Using a non-parametric Bayesian method allows per-pixel mode counts to be automatically inferred, avoiding over-/under- fitting. We also develop novel model learning algorithms for continuous update of the model in a principled fashion as the scene changes. These key advantages enable us to outperform the state-of-the-art alternatives on four benchmarks.},
number = {4},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
author = {Haines, T.S.F. and Xiang, Tao},
month = {April},
year = {2014},
note = {00005},
keywords = {background},
pages = {670--683}
}
Downloads: 0
{"_id":"6G8jmh9vvebdR9Tvo","bibbaseid":"haines-xiang-backgroundsubtractionwithdirichletprocessmixturemodels-2014","downloads":0,"creationDate":"2015-09-03T07:18:49.184Z","title":"Background Subtraction with Dirichlet Process Mixture Models","author_short":["Haines, T.","Xiang, T."],"year":2014,"bibtype":"article","biburl":"http://bibbase.org/zotero/fred.qi","bibdata":{"abstract":"Video analysis often begins with background subtraction. This problem is often approached in two steps-a background model followed by a regularisation scheme. A model of the background allows it to be distinguished on a per-pixel basis from the foreground, whilst the regularisation combines information from adjacent pixels. We present a new method based on Dirichlet process Gaussian mixture models, which are used to estimate per-pixel background distributions. It is followed by probabilistic regularisation. Using a non-parametric Bayesian method allows per-pixel mode counts to be automatically inferred, avoiding over-/under- fitting. We also develop novel model learning algorithms for continuous update of the model in a principled fashion as the scene changes. These key advantages enable us to outperform the state-of-the-art alternatives on four benchmarks.","author":["Haines, T.S.F.","Xiang, Tao"],"author_short":["Haines, T.","Xiang, T."],"bibtex":"@article{ haines_background_2014,\n title = {Background {Subtraction} with {Dirichlet} {Process} {Mixture} {Models}},\n volume = {36},\n issn = {0162-8828},\n doi = {10.1109/TPAMI.2013.239},\n abstract = {Video analysis often begins with background subtraction. This problem is often approached in two steps-a background model followed by a regularisation scheme. A model of the background allows it to be distinguished on a per-pixel basis from the foreground, whilst the regularisation combines information from adjacent pixels. We present a new method based on Dirichlet process Gaussian mixture models, which are used to estimate per-pixel background distributions. It is followed by probabilistic regularisation. Using a non-parametric Bayesian method allows per-pixel mode counts to be automatically inferred, avoiding over-/under- fitting. We also develop novel model learning algorithms for continuous update of the model in a principled fashion as the scene changes. These key advantages enable us to outperform the state-of-the-art alternatives on four benchmarks.},\n number = {4},\n journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},\n author = {Haines, T.S.F. and Xiang, Tao},\n month = {April},\n year = {2014},\n note = {00005},\n keywords = {background},\n pages = {670--683}\n}","bibtype":"article","doi":"10.1109/TPAMI.2013.239","id":"haines_background_2014","issn":"0162-8828","journal":"IEEE Transactions on Pattern Analysis and Machine Intelligence","key":"haines_background_2014","keywords":"background","month":"April","note":"00005","number":"4","pages":"670--683","title":"Background Subtraction with Dirichlet Process Mixture Models","type":"article","volume":"36","year":"2014","bibbaseid":"haines-xiang-backgroundsubtractionwithdirichletprocessmixturemodels-2014","role":"author","urls":{},"keyword":["background"],"downloads":0},"search_terms":["background","subtraction","dirichlet","process","mixture","models","haines","xiang"],"keywords":["background"],"authorIDs":[],"dataSources":["y8en6y5RHukfeLuPH"]}