A Lagrangian Particle Dynamics Approach for Crowd Flow Segmentation and Stability Analysis

This paper proposes a framework in which Lagrangian particle dynamics is used for the segmentation of high density crowd flows and detection of flow instabilities. For this purpose, a flow field generated by a moving crowd is treated as an aperiodic dynamical system. A grid of particles is overlaid on the flow field, and is advected using a numerical integration scheme. The evolution of particles through the flow is tracked using a flow map, whose spatial gradients are subsequently used to setup a Cauchy Green deformation tensor for quantifying the amount by which the neighboring particles have diverged over the length of the integration. The maximum eigenvalue of the tensor is used to construct a finite time Lyapunov exponent (FTLE) field, which reveals the Lagrangian coherent structures (LCS) present in the underlying flow. The LCS divide flow into regions of qualitatively different dynamics and are used to locate boundaries of the flow segments in a normalized cuts framework. Any change in the number of flow segments over time is regarded as an instability, which is detected by establishing correspondences between flow segments over time. The experiments are conducted on a challenging set of videos taken from Google Video and a National Geographic documentary.

[1]  Jitendra Malik,et al.  Normalized Cuts and Image Segmentation , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  G. Haller.  Finding finite-time invariant manifolds in two-dimensional velocity fields. , 2000, Chaos.

[3]  J. Marsden,et al.  Tricubic interpolation in three dimensions , 2005 .

[4]  Nuno Vasconcelos,et al.  Mixtures of dynamic textures , 2005, Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1.

[5]  J. Marsden,et al.  Definition and properties of Lagrangian coherent structures from finite-time Lyapunov exponents in two-dimensional aperiodic flows , 2005 .

[6]  G. Haller,et al.  Lagrangian coherent structures and mixing in two-dimensional turbulence , 2000 .

[7]  P. Reisman,et al.  Crowd detection in video sequences , 2004, IEEE Intelligent Vehicles Symposium, 2004.

[8]  Serge J. Belongie,et al.  Counting Crowded Moving Objects , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[9]  Jitendra Malik,et al.  Normalized cuts and image segmentation , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[10]  H. Abarbanel,et al.  Lyapunov exponents from observed time series. , 1990, Physical review letters.

[11]  Ramakant Nevatia,et al.  Bayesian human segmentation in crowded situations , 2003, 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings..

[12]  Peter H. Tu,et al.  Crowd Segmentation Through Emergent Labeling , 2004, ECCV Workshop SMVP.

[13]  Roberto Cipolla,et al.  Unsupervised Bayesian Detection of Independent Motion in Crowds , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[14]  Leonidas J. Guibas,et al.  Counting people in crowds with a real-time network of simple image sensors , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[15]  G. Haller.  Distinguished material surfaces and coherent structures in three-dimensional fluid flows , 2001 .

[16]  G. Lapeyre,et al.  Characterization of finite-time Lyapunov exponents and vectors in two-dimensional turbulence. , 2002, Chaos.

[17]  George Haller,et al.  The geometry and statistics of mixing in aperiodic flows , 1999 .

[18]  Seth J. Teller,et al.  Particle Video: Long-Range Motion Estimation Using Point Trajectories , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).