Earticle

초록

영어

This paper focuses on the detection of the abnormal motion behaviour recognition of the crowd, and proposes an innovation method which is consist of three steps, i.e. SIFT flow + weighted orientation histogram + Hidden Markov Model(HMM). Analogous to optical flow, which is used to get the motion information of the pixels from two adjacent frames, SIFT flow is of higher precision. Next, we build up a a weighted orientation histogram as a statistical measurement for the SIFT flow features from the first step. Finally, the derived histogram is taken as the input for HMM in preparation for the detection of abnormal crowd motion. Experimental results show that compared to the existing method, our proposed one can detect the abnormal motion behaviour more effectively.

목차

Abstract
 1. Introduction
 2. SIFT Flow Technology
  2.1 Dense SIFT Descriptors
  2.2 SIFT Flow Flied
  2.3 Optimization of SIFT Flow Flied
  2.4 Weighted Orientation Histogram
 3. Hidden Markov Model
 4. Experimental Result
  4.1 SIFT Flow Processing
  4.2 HMM Processing
  4.3 Abnormal Behaviour Detection
 5. Conclusions
 References

키워드

저자

• Dongping Zhang [ College of Information Engineering China Jiliang University, Hangzhou, 310018, China ]
• Kaihang Xu [ College of Information Engineering China Jiliang University, Hangzhou, 310018, China ]
• Huailiang Peng [ College of Information Engineering China Jiliang University, Hangzhou, 310018, China ]
• Ye Shen [ College of Information Engineering China Jiliang University, Hangzhou, 310018, China ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Signal Processing, Image Processing and Pattern Recognition

• 간기

  격월간

• pISSN

  2005-4254

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605