A shape derivative based approach for crowd flow segmentation

  • Authors:
  • Si Wu;Zhiwen Yu;Hau-San Wong

  • Affiliations:
  • Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong;Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong;Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong

  • Venue:
  • ACCV'09 Proceedings of the 9th Asian conference on Computer Vision - Volume Part I
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Crowd movement analysis has many practical applications, especially for video surveillance. The common methods are based on pedestrian detection and tracking. With an increase of crowd density, however, it is difficult for these methods to analyze crowd movement because of the computation and complexity. In this paper, a novel approach for crowd flow segmentation is proposed. We employ a Weighting Fuzzy C-Means clustering algorithm (WFCM) to extract the motion region in optical flow field. In order to further analyze crowd movement, we make use of translation flow to approximate local crowd movement, and design a shape derivative based region growing scheme to segment the crowd flows. In the experiments, the proposed method is tested on a set of crowd video sequences from low density to high density.