Motion Flow Segmentation Based on Optical Flow Angle

DOI

10.2991/icwcsn-16.2017.120How to use a DOI?

Keywords

Optical flow; Lyapunov exponent; Crowd target detection; Motion flow segmentation.

Abstract

Aiming at the irrationality of deep motion state analysis, such as density estimation, target tracking and behavior understanding, which taken the motion foreground as a whole in dense crowd scene. This paper proposes an algorithm to segment motion flow based on optical flow angle. First of all, the motion prospect is obtained by using the finite-time Lyapunov exponent (FTLE) algorithm. Then get the foreground optical flow information with the prospect as a mask. Finally, the dynamic K-means clustering method is applied to segment the foreground optical flow angle. Experimental results show that the proposed algorithm is clear and easy to implement, and the motion flow can be segmented accurately in different scenes.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Kun Yang
AU  - Lin Wang
AU  - Fu-Jian Feng
AU  - Jiang-Hao Yu
AU  - Yuan-Fei Cheng
PY  - 2016/12
DA  - 2016/12
TI  - Motion Flow Segmentation Based on Optical Flow Angle
BT  - Proceedings of the 3rd International Conference on Wireless Communication and Sensor Networks (WCSN 2016)
PB  - Atlantis Press
SP  - 592
EP  - 597
SN  - 2352-538X
UR  - https://doi.org/10.2991/icwcsn-16.2017.120
DO  - 10.2991/icwcsn-16.2017.120
ID  - Yang2016/12
ER  -