SEPARATING BACKGROUND AND FOREGROUND IN VIDEO BASED ON A NONPARAMETRIC BAYESIAN MODEL

Xinghao Ding; Lawrence Carin

SEPARATING BACKGROUND AND FOREGROUND IN VIDEO BASED ON A NONPARAMETRIC BAYESIAN MODEL

Xinghao Ding^*, Lawrence Carin

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Separating background and foreground in video is a fundamental problem in computer vision. We present a Bayesian hierarchical model to address this challenge, and apply it to video with dynamic scenes. The model uses a nonparametric prior, a beta-bernoulli process, for both the background and foreground representation. Additionally, the model uses neighborhood information of each pixel to encourage group clustering of the foreground. A collapsed Gibbs sampler is used for efficient posterior inference. Experimental results show competitive performance of the proposed model.

Original language	English
Title of host publication	2011 IEEE STATISTICAL SIGNAL PROCESSING WORKSHOP (SSP)
Publisher	IEEE
Pages	321-324
Number of pages	4
State	Published - 2011
Externally published	Yes
Event	IEEE Statistical Signal Processing Workshop (SSP) - Nice, France Duration: Jun 28 2011 → Jun 30 2011

Conference

Conference	IEEE Statistical Signal Processing Workshop (SSP)
Country/Territory	France
City	Nice
Period	06/28/11 → 06/30/11

Keywords

Background subtraction
dynamic scenes
nonparametric Bayesian hierarchical model
beta-bernoulli process
group sparsity

Cite this

@inproceedings{5fbf46a88c7e471aaae5a86873c009b8,

title = "SEPARATING BACKGROUND AND FOREGROUND IN VIDEO BASED ON A NONPARAMETRIC BAYESIAN MODEL",

abstract = "Separating background and foreground in video is a fundamental problem in computer vision. We present a Bayesian hierarchical model to address this challenge, and apply it to video with dynamic scenes. The model uses a nonparametric prior, a beta-bernoulli process, for both the background and foreground representation. Additionally, the model uses neighborhood information of each pixel to encourage group clustering of the foreground. A collapsed Gibbs sampler is used for efficient posterior inference. Experimental results show competitive performance of the proposed model.",

keywords = "Background subtraction, dynamic scenes, nonparametric Bayesian hierarchical model, beta-bernoulli process, group sparsity",

author = "Xinghao Ding and Lawrence Carin",

year = "2011",

language = "English",

pages = "321--324",

booktitle = "2011 IEEE STATISTICAL SIGNAL PROCESSING WORKSHOP (SSP)",

publisher = "IEEE",

note = "IEEE Statistical Signal Processing Workshop (SSP) ; Conference date: 28-06-2011 Through 30-06-2011",

}

TY - GEN

T1 - SEPARATING BACKGROUND AND FOREGROUND IN VIDEO BASED ON A NONPARAMETRIC BAYESIAN MODEL

AU - Ding, Xinghao

AU - Carin, Lawrence

PY - 2011

Y1 - 2011

N2 - Separating background and foreground in video is a fundamental problem in computer vision. We present a Bayesian hierarchical model to address this challenge, and apply it to video with dynamic scenes. The model uses a nonparametric prior, a beta-bernoulli process, for both the background and foreground representation. Additionally, the model uses neighborhood information of each pixel to encourage group clustering of the foreground. A collapsed Gibbs sampler is used for efficient posterior inference. Experimental results show competitive performance of the proposed model.

AB - Separating background and foreground in video is a fundamental problem in computer vision. We present a Bayesian hierarchical model to address this challenge, and apply it to video with dynamic scenes. The model uses a nonparametric prior, a beta-bernoulli process, for both the background and foreground representation. Additionally, the model uses neighborhood information of each pixel to encourage group clustering of the foreground. A collapsed Gibbs sampler is used for efficient posterior inference. Experimental results show competitive performance of the proposed model.

KW - Background subtraction

KW - dynamic scenes

KW - nonparametric Bayesian hierarchical model

KW - beta-bernoulli process

KW - group sparsity

M3 - Conference contribution

SP - 321

EP - 324

BT - 2011 IEEE STATISTICAL SIGNAL PROCESSING WORKSHOP (SSP)

PB - IEEE

T2 - IEEE Statistical Signal Processing Workshop (SSP)

Y2 - 28 June 2011 through 30 June 2011

ER -

SEPARATING BACKGROUND AND FOREGROUND IN VIDEO BASED ON A NONPARAMETRIC BAYESIAN MODEL

Abstract

Conference

Keywords

Fingerprint

Cite this