배경 분리 알고리즘 기반 이동 객체 탐지 성능 평가 기법 연구 = Performance Evaluation of Background Subtraction Based Moving Object Detection Approach

다국어 초록 (Multilingual Abstract)

The background subtraction technique finds moving objects and reconstructs the background from video sequences. The background subtraction has extensive real-world applications.
Most of the background subtraction studies have focused on increasing the accuracy while reducing the complexity. Though few studies have appraised the accuracy of the background subtraction methods, the researchers have not measured the computational complexity of the methods. Thus, in this study, our main goal was to measure the accuracy and computational complexity of the background subtraction approaches. This study can be used in industry and academy. Also, we implemented and assessed the performance of the three different types of background subtraction algorithms such as the cluster-based method, the statistical-based method, and the sample consensusbased method. We mainly used the F-measure with other confusion metrics, which are the most accepted criteria to assess the segmentation accuracy of the background subtraction algorithms. Also, we evaluated the complexity in terms of the memory usage per pixel and the number of frame display per second for the CDD-2012, CDD-2014, and Carnegie Mellon datasets. The experimental data are presented in the table in Section 4 to show the accuracy and computational complexity.

국문 초록 (Abstract)

배경 분리 알고리즘은 비디오 시퀀스에서 배경을 재구성하여 움직이는 객체를 찾아내기 위한 기술로, 광범위한 응용 분야에 활용되고 있다. 배경 분리 알고리즘의 연구는 주로 복잡도를 줄이면서 정확성을 높이는 데 중점을 두고 있으나, 복잡도를 측정하거나 정확도를 평가하는 방법에 대한 연구는 상대적으로 미비한 상태이다. 따라서, 본 연구에서는 산업 및 학술적으로 모두 사용 가능한 배경 분리 알고리즘의 정확도와 계산 복잡도 평가 방안을 제시하고자 한다. 본 논문에서는 클러스터 기반 기법, 통계 기반 기법, 표본 합의 기반 기법의 세가지 종류와 배경 분리 알고리즘을 구현하고 평가한다. 특히 배경 분리 알고리즘의 분할 정확도를 평가하는데 가장 적합한 방법인 F-measure를 다른 혼합 지표와 함께 사용하였으며, CDD-2012, CDD-2014, 카네기멜론의 데이터시트를 픽셀당 메모리 사용량, 초당 프레임 수로 복잡도를 평가하여 정확도와 계산 복잡도를 나타낸 실험 결과를 본문(섹션 4)에 제시하였다.

참고문헌 (Reference)

1 A. B. Godbehere, "Visual tracking of human visitors under variablelighting conditions for a responsive audio art install ation" 4305-4312, 2012

2 P. KaewTraKulPong, "Video-Based Surveillance Systems: Computer Vision and Distributed Processing" Springer US 135-144, 2002

3 D. K. Prasad, "Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment : A Survey" 18 (18): 1993-2016, 2017

4 O. Barnich, "ViBe : A Universal Background Subtraction Algorithm for Video Sequences" 20 (20): 1709-1724, 2011

5 W. E. L. Grimson, "Using adaptive tracking to classify and monitor activities in a site" 22-29, 1998

6 P. -L. St-Charles, "Universal Background Subtraction Using Word Consensus Models" 25 (25): 4768-4781, 2016

7 P. -L. St-Charles, "SuBSENSE : A Universal Change Detection Method With Local Adaptive Sensitivity" 24 (24): 359-373, 2015

8 D. Ortego, "Stand-alone quality estimation of background subtraction algorithms" 162 : 87-102, 2017

9 K. Kim, "Real-time foreground-background segmentation using codebook model" 11 (11): 172-185, 2004

10 P. Sarisaray-Boluk, "Performance evaluation of background subtraction algorithms for Android devices deployed in Wireless Multimedia Sensor Networks" 779-784, 2014

11 M. A. Hossain, "Performance Assessment of Background Subtraction Algorithm" 1049-1051, 2019

12 S. Manipriya, "Performance Analysis of Spatial Color Information for Object Detection Using Background Subtraction" 10 : 63-69, 2014

13 Kim, T, "PDR: A Performance Evaluation Method for Foreground-Background Segmentation Algorithms" 2006

14 "OpenCV: cv::BackgroundSubtractorMOG2 Class Reference"

15 J. C. SanMiguel, "On the Evaluation of Background Subtraction Algorithms without Ground-Truth" 180-187, 2010

16 D. K. Prasad, "Object Detection in a Maritime Environment : Performance Evaluation of Background Subtraction Methods" 20 (20): 1787-1802, 2019

17 A. Elgammal, "Nonparametric Model for Background Subtraction" 751-767, 2000

18 C. Stauffer, "Learning patterns of activity using real-time tracking" 22 (22): 747-757, 2000

19 P. -L. St-Charles, "Improving background subtraction using Local Binary Similarity Patterns" 509-515, 2014

20 Z. Zivkovic, "Improved adaptive Gaussian mixture model for background subtraction" ICPR 2004 2004 : 28-31, 2004

21 S. Brutzer, "Evaluation of background subtraction techniques for video surveillance" 1937-1944, 2011

22 J. Sepulveda, "Evaluation of background subtraction algorithms using MuHAVi, a multicamera human action video dataset" 10-14, 2014

23 Z. Zivkovic, "Efficient adaptive density estimation per image pixel for the task of background subtraction" 27 (27): 773-780, 2005

24 N. Goyette, "Changedetection. net : A new change detection benchmark dataset" 1-8, 2012

25 Y. Wang, "CDnet 2014: An Expanded Change Detection Benchmark Dataset" 393-400, 2014

26 M. Van Droogenbroeck, "Background subtraction : Experiments and improvements for ViBe" 32-37, 2012

27 M. Hofmann, "Background segmentation with feedback : The Pixel-Based Adaptive Segmenter" 38-43, 2012

28 T. Bouwmans, "Background Subtraction in Real Applications: Challenges, Current Models and Future Directions" ArXiv190103577Cs

29 S. K. Choudhury, "An Evaluation of Background Subtraction for Object Detection Vis-a-Vis Mitigating Challenging Scenarios" 4 : 6133-6150, 2016

30 C. Stauffer, "Adaptive background mixture models for real-time tracking" 2 : 246-252, 1999