Gaussian mixture models-based background subtraction: Improvements and application in surveillance = 가우시안 혼합 모델에 기반한 배경 제거 기법: 영상 감시 분야에서의 개선 및 응용

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 503
  • Download : 0
Surveillance systems are finding widespread applications and have attracted a lot interest from industry and academia. In this work, event-based surveillance video recording or transmission is discussed. Motion is described as the event of interest and Gaussian mixture models-based background subtraction technique is used to detect moving objects in a given scene. In Chapter 2, a background subtraction scheme based on Gaussian Mixture Models (GMM) is proposed where the speed of motion detection is increased and the number of false positives is reduced when compared to conventional GMM-based foreground detection schemes, by suppressing excessive update of mean and variance values. A condition is added to determine if the change in the mean and variance of the background model is greater than a specified threshold; the mean and variance are updated only if the condition is satisfied. As a result, 35% increase in speed was observed with test sequences on a PC with the false positive rate decreased from 6% to 0.3%. Experiment with live video on an ARM-based single board computer showed an increase in speed of background subtraction from 12.5fps to 20fps with the proposed method. The proposed method is proven to be more robust and suitable for use especially in low power video surveillance applications compared to the original GMM-based background subtraction schemes. CCTV-based surveillance systems gaining widespread popularity still waste computational power, transmission bandwidth and storage space. Chapter 3 tries to respond to this necessity by proposing a motion-based video recording scheme with dual frame rate motion detection. Statistical models for memory and energy consumption of the overall system are described. Root-mean-square error of the models for memory consumption is 0.54 and 0.78 for systems with single frame rate and dual frame rate motion detection, respectively. For a typical surveillance video, the proposed dual frame rate system stores 1...
Advisors
Kyung, Chong-Minresearcher경종민
Description
한국과학기술원 : 전기및전자공학과,
Publisher
한국과학기술원
Issue Date
2014
Identifier
592379/325007  / 020104504
Language
eng
Description

학위논문(석사) - 한국과학기술원 : 전기및전자공학과, 2014.8, [ v, 36 p. ]

Keywords

Gaussian Mixture Model; 저 에너지; 저 메모리; 사건 감지; 동작 분석; 혼합 모델; Image Motion analysis; Event Detection; Low Memory; Low Energy

URI
http://hdl.handle.net/10203/196652
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=592379&flag=dissertation
Appears in Collection
EE-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0