A Lagrangian framework for video analytics

Abstract

The extraction of motion patterns from image sequences based on the optical flow methodology is an important and timely topic among visual multi media applications. In this work we will present a novel framework that combines the optical flow methodology from image processing with methods developed for the Lagrangian analysis of time-dependent vector fields. The Lagrangian approach has been proven to be a valuable and powerful tool to capture the complex dynamic motion behavior within unsteady vector fields. To come up with a compact and applicable framework, this paper will provide concepts on how to compute trajectory-based Lagrangian measures in series of optical flow fields, a set of basic measures to capture the essence of the motion behavior within the image, and a compact hierarchical, featurebased description of the resulting motion features. The resulting framework will bee shown to be suitable for an automated image analysis as well as compact visual analysis of image sequences in its spatio-temporal context. We show its applicability for the task of motion feature description and extraction on different temporal scales, crowd motion analysis, and automated detection of abnormal events within video sequences.

Video-based Bottleneck Detection utilizing Lagrangian Dynamics in Crowded Scenes

“…the crowd is assessed as a single entity. The behaviors of individuals in a crowd are dependent on the crowd behavior [17,9] and modelled by fluid dynamic processes [2,23,11,8]. Hughes work [7] supports the assumption that crowds are a flowing continuum and proposed three main behavioral hypotheses for persons moving in a crowd.…”

“…The idea of this detector is that physical bottlenecks are related to bottlenecks in the contours of the crowd flow segments. To segment the crowd flow contour we apply long-term analysis based on the Lagrangian framework proposed in [8] and use the Finite Time Lyapunov Exponents (FTLE) field to extract motion boundaries. High ridges in the FTLE field indicate Lagrangian features that are assumed to be located at motion boundaries.…”

Super-pixel based crowd flow segmentation in H.264 compressed videos

“…Anil et al [12] had outlined the review of various computer vision algorithms dealing with crowded scenarios and proposed a system for the automatic detection of dominant patterns of crowd flow in dense crowd scenarios by tracking the low level object features using the optical flow algorithm. Kuhn et al [13] had developed a frame work for extracting the motion patterns by combining the optical flow with Lagrangian analysis of time dependent vector fields shown its applicability for crowd analysis like automated detection of abnormal events in the video sequences. Another significant work to address this problem was done by Ali et al [1], where they have modeled this problem from the perspective of fluid dynamics.…”

Crowd flow segmentation based on motion vectors in H.264 compressed domain

“…Another work was done by Li et al [5] which extracts the dynamic region of the crowded scenes and the crowd segmentation is performed based on the histogram curve of the angle information of the foreground velocity field. Kuhn et al [6] had developed a frame work for extracting the motion patterns by combining the optical flow from image processing with lagrangian analysis of time dependent vector fields. They had shown its applicability for crowd analysis like automated detection of abnormal events in the video sequences.…”