한국전자통신학회논문지 (The Journal of the Korea institute of electronic communication sciences)

한국전자통신학회 (Korea Institute of Electronic Communication Science)
권호 표지
DOI QR코드

DOI QR Code

움직임 벡터와 GPU를 이용한 인간 활동성 분석

Analysis of Human Activity Using Motion Vector and GPU

  • 김선우 (군산대학교 정보통신공학과) ;
  • 최연성 (군산대학교 정보통신공학과)
  • 투고 : 2014.08.12
  • 심사 : 2014.10.17
  • 발행 : 2014.10.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

본 논문에서는 실시간 감시 시스템에서 인간의 활동성을 분석하기 위하여 움직임 벡터를 사용하며, 고속연산에 GPU를 활용한다. 먼저 가장 중요한 부분인 전경으로부터 적응적 가우시안 혼합기법, 두드러진 움직임을 위한 가중치 차영상 기법, 움직임 벡터를 이용하여 인간이라고 판단되는 블랍을 검출하고, 추출된 움직임 벡터를 이용하여 사람의 활동성을 분석한다. 본 논문에서는 사람의 행동을 크게 {Active, Inactive}, {Position Moving, Fixed Moving}, {Walking, Running}의 세 가지 메타 클래스로 분류하고 인식하였다. 실험을 위해서 약 300개의 상황을 연출하였으며, 약 86%~98% 의 인식률을 보였다. 또한 $1920{\times}1080$ 크기 영상에서 CPU 기반은 4.2초 정도 걸렸는데, GPU 기반에서는 0.4초 이내로 빨라진 결과를 얻었다.

In this paper, We proposed the approach of GPU and motion vector to analysis the Human activity in real-time surveillance system. The most important part, that is detect blob(human) in the foreground. We use to detect Adaptive Gaussian Mixture, Weighted subtraction image for salient motion and motion vector. And then, We use motion vector for human activity analysis. In this paper, the activities of human recognize and classified such as meta-classes like this {Active, Inactive}, {Position Moving, Fixed Moving}, {Walking, Running}. We created approximately 300 conditions for the simulation. As a result, We showed a high success rate about 86~98%. The results also showed that the high resolution experiment by the proposed GPU-based method was over 10 times faster than the cpu-based method.

키워드

참고문헌

  1. S.-W. Kim, T.-R. Ha, C.-B. Park, and Y.-S. Choi, "Salient Motion Information Detection Method Using Weighted Subtraction Image and Motion Vector," J. of the Korean Institute of Maritime Information and Communication Sciences, vol. 11, no. 4, 2007. pp. 779-785.
  2. J.-J. Park, S.-W. Kim, Y.-S. Choi, C.-B. Park, and T.-R. Ha, "A Study On the Moving Object Tracking System Using Multi-feature Matching," J. of the Korean Institute of Maritime Information and Communication Sciences, vol. 11, no. 4, 2007, pp. 787-792.
  3. H.-T. Kim, G.-H. Lee, J.-S. Park, and Y.-S. Yu, "Vehicle Detection in Tunnel using Gaussian Mixture Model and Mathematical Morphological Processing," J. of the Korea Institute of Electronic Communication Sciences, vol. 7, no. 5, 2012, pp. 967-974.
  4. S.-H. Lee, "Fast motion Estimation with Adaptive Search Range Adjustment using Motion Activities of Temporal and Spatial Neighbor Blocks," J. of the Korea Institute of Electronic Communication Sciences, vol. 5, no. 4, 2010, pp. 372-378.
  5. S.-W. Kim, Y.-S. Choi, and H.-K. Yang, "Analysis of Human Activity Using Silhouette and Feature Parameters," J. of the Korean Institute of Maritime Information and Communication Sciences, vol. 15, no. 2, 2011, pp. 923-926.
  6. S.-W. Kim, Y.-S. Choi, and H.-K. Yang, "Analysis of Human Activity Using motion Vector," J. of the Korean Institute of Maritime Information and Communication Sciences, vol. 15, no. 2, 2011, pp. 157-160. https://doi.org/10.6109/jkiice.2011.15.1.157
  7. C. Stauffer and W. Grimson, "Adaptive background mixture models for real-time tracking," In CVPR 99, Fort Collins, FL, June vol. 2, 1999, pp. 246-252.
  8. Z. Zivkovic and F. van der Heijden, "Efficient adaptive density estimation per image pixel for the task of background subtraction," Pattern Recognition Letters, vol. 27, no. 7, 2006, pp. 773-780. https://doi.org/10.1016/j.patrec.2005.11.005
  9. J. Sanders and E. Kandrot, CUDA by Example : An Introdution to General Purpose GPU Programming. Boston : Addison Wesley, 2010.