Fast illumination-robust foreground detection using hierarchical distribution map for real-time video surveillance system

Jongin Son, Seungryong Kim, Kwanghoon Sohn

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Foreground detection is one of the most important and fundamental tasks in many computer vision applications such as real-time video surveillance. Although there have been many efforts to find solutions to this problem, many obstacles such as illumination changes, noises, dynamic backgrounds, and computational complexities have prevented them from being used in real surveillance systems. In this paper, to alleviate these inherent limitations of conventional methods, we propose a fast illumination-robust foreground detection (FIFD) system that provides robustness against illumination variations and noises from various real circumstances with an efficient computational scheme. In contrast to the conventional approaches, our method focuses on efficiently formulating the foreground object detection system by leveraging a foreground candidate region detection and hierarchical distribution map. Specifically, our approach consists of three parts. First, for a query image, foreground candidates are detected by fusing multiple methods. The existence and the block size of the foreground object are determined through the use of the foreground continuity. Second, the foreground block is found from the estimated distribution map and then detected from the extracted valid blocks. Finally, with a labeling scheme, the foreground is detected. To intensively evaluate our approach compared to the conventional methods, we use the publicly available I2R and traffic datasets, and we build a novel electron multiplying charge-coupled device foreground detection benchmark taken in an environment with light lower than 10lux. Experimental results show that our approach provides satisfactory performance compared to the state-of-the-art methods even under very challenging circumstances. Furthermore, our approach is very efficient in that it takes only approximately 31 ms per frame.

Original languageEnglish
Pages (from-to)32-41
Number of pages10
JournalExpert Systems with Applications
Volume66
DOIs
Publication statusPublished - 2016 Dec 30

Fingerprint

Lighting
Charge coupled devices
Labeling
Computer vision
Computational complexity
Electrons
Object detection

All Science Journal Classification (ASJC) codes

  • Engineering(all)
  • Computer Science Applications
  • Artificial Intelligence

Cite this

@article{4ad0cb739f524e5db8a9ac75ce730d9b,
title = "Fast illumination-robust foreground detection using hierarchical distribution map for real-time video surveillance system",
abstract = "Foreground detection is one of the most important and fundamental tasks in many computer vision applications such as real-time video surveillance. Although there have been many efforts to find solutions to this problem, many obstacles such as illumination changes, noises, dynamic backgrounds, and computational complexities have prevented them from being used in real surveillance systems. In this paper, to alleviate these inherent limitations of conventional methods, we propose a fast illumination-robust foreground detection (FIFD) system that provides robustness against illumination variations and noises from various real circumstances with an efficient computational scheme. In contrast to the conventional approaches, our method focuses on efficiently formulating the foreground object detection system by leveraging a foreground candidate region detection and hierarchical distribution map. Specifically, our approach consists of three parts. First, for a query image, foreground candidates are detected by fusing multiple methods. The existence and the block size of the foreground object are determined through the use of the foreground continuity. Second, the foreground block is found from the estimated distribution map and then detected from the extracted valid blocks. Finally, with a labeling scheme, the foreground is detected. To intensively evaluate our approach compared to the conventional methods, we use the publicly available I2R and traffic datasets, and we build a novel electron multiplying charge-coupled device foreground detection benchmark taken in an environment with light lower than 10lux. Experimental results show that our approach provides satisfactory performance compared to the state-of-the-art methods even under very challenging circumstances. Furthermore, our approach is very efficient in that it takes only approximately 31 ms per frame.",
author = "Jongin Son and Seungryong Kim and Kwanghoon Sohn",
year = "2016",
month = "12",
day = "30",
doi = "10.1016/j.eswa.2016.08.062",
language = "English",
volume = "66",
pages = "32--41",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier Limited",

}

Fast illumination-robust foreground detection using hierarchical distribution map for real-time video surveillance system. / Son, Jongin; Kim, Seungryong; Sohn, Kwanghoon.

In: Expert Systems with Applications, Vol. 66, 30.12.2016, p. 32-41.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Fast illumination-robust foreground detection using hierarchical distribution map for real-time video surveillance system

AU - Son, Jongin

AU - Kim, Seungryong

AU - Sohn, Kwanghoon

PY - 2016/12/30

Y1 - 2016/12/30

N2 - Foreground detection is one of the most important and fundamental tasks in many computer vision applications such as real-time video surveillance. Although there have been many efforts to find solutions to this problem, many obstacles such as illumination changes, noises, dynamic backgrounds, and computational complexities have prevented them from being used in real surveillance systems. In this paper, to alleviate these inherent limitations of conventional methods, we propose a fast illumination-robust foreground detection (FIFD) system that provides robustness against illumination variations and noises from various real circumstances with an efficient computational scheme. In contrast to the conventional approaches, our method focuses on efficiently formulating the foreground object detection system by leveraging a foreground candidate region detection and hierarchical distribution map. Specifically, our approach consists of three parts. First, for a query image, foreground candidates are detected by fusing multiple methods. The existence and the block size of the foreground object are determined through the use of the foreground continuity. Second, the foreground block is found from the estimated distribution map and then detected from the extracted valid blocks. Finally, with a labeling scheme, the foreground is detected. To intensively evaluate our approach compared to the conventional methods, we use the publicly available I2R and traffic datasets, and we build a novel electron multiplying charge-coupled device foreground detection benchmark taken in an environment with light lower than 10lux. Experimental results show that our approach provides satisfactory performance compared to the state-of-the-art methods even under very challenging circumstances. Furthermore, our approach is very efficient in that it takes only approximately 31 ms per frame.

AB - Foreground detection is one of the most important and fundamental tasks in many computer vision applications such as real-time video surveillance. Although there have been many efforts to find solutions to this problem, many obstacles such as illumination changes, noises, dynamic backgrounds, and computational complexities have prevented them from being used in real surveillance systems. In this paper, to alleviate these inherent limitations of conventional methods, we propose a fast illumination-robust foreground detection (FIFD) system that provides robustness against illumination variations and noises from various real circumstances with an efficient computational scheme. In contrast to the conventional approaches, our method focuses on efficiently formulating the foreground object detection system by leveraging a foreground candidate region detection and hierarchical distribution map. Specifically, our approach consists of three parts. First, for a query image, foreground candidates are detected by fusing multiple methods. The existence and the block size of the foreground object are determined through the use of the foreground continuity. Second, the foreground block is found from the estimated distribution map and then detected from the extracted valid blocks. Finally, with a labeling scheme, the foreground is detected. To intensively evaluate our approach compared to the conventional methods, we use the publicly available I2R and traffic datasets, and we build a novel electron multiplying charge-coupled device foreground detection benchmark taken in an environment with light lower than 10lux. Experimental results show that our approach provides satisfactory performance compared to the state-of-the-art methods even under very challenging circumstances. Furthermore, our approach is very efficient in that it takes only approximately 31 ms per frame.

UR - http://www.scopus.com/inward/record.url?scp=84986261704&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986261704&partnerID=8YFLogxK

U2 - 10.1016/j.eswa.2016.08.062

DO - 10.1016/j.eswa.2016.08.062

M3 - Article

AN - SCOPUS:84986261704

VL - 66

SP - 32

EP - 41

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

ER -