Human Tracking Method Based on Maximally Stable Extremal Regions with Multi-Cameras

With respect to the human tracking with multi-cameras in the video surveillance system, a human tracking method based on MSER (Maximally Stable Extremal Regions) was established. The approach transforms the human tracking into elliptic region matching. The method does elliptic region fitting to each MSER, and then selects the elliptic regions which meet some constraints. These selected elliptic regions are normalized to unity circular regions. The right matched elliptic regions are gotten by rotational invariant vectors calculation, histogram density estimation and weighted average distance calculation. Experimental results show that the approach can effectively realize the human tracking with multi-cameras.

LAX SHOCKS IN MIXED-TYPE SYSTEMS OF CONSERVATION LAWS

Small amplitude shocks involving a state with complex characteristic speeds arise in mixed-type systems of two or more conservation laws. We study such shocks in detail in the generic case, when they appear near the codimension-1 elliptic boundary. Then we classify all exceptional codimension-2 states on smooth parts of the elliptic boundary. Asymptotic formulae describing shock curves near regular and exceptional states are derived. The type of singularity at the exceptional point depends on the second and third derivatives of the flux function. The main application is understanding the structure of small amplitude Riemann solutions where one of the initial states lies in the elliptic region.

THE ELLIPTICITY PRINCIPLE FOR SELF-SIMILAR POTENTIAL FLOWS

We consider self-similar potential flow for compressible gas with polytropic pressure law. Self-similar solutions arise as large-time asymptotes of general solutions, and as exact solutions of many important special cases like Mach reflection, multidimensional Riemann problems, or flow around corners. Self-similar potential flow is a quasilinear second-order PDE of mixed type which is hyperbolic at infinity (if the velocity is globally bounded). The type in each point is determined by the local pseudo-Mach number L, with L < 1 (respectively, L > 1) corresponding to elliptic (respectively, hyperbolic) regions. We prove an ellipticity principle: the interior of a parabolic-elliptic region of a sufficiently smooth solution must be elliptic; in fact L must be bounded above away from 1 by a domain-dependent function. In particular there are no open parabolic regions. We also discuss the case of slip boundary conditions at straight solid walls.

THE ROLE OF RELATIONAL CONSTRAINTS IN REGION MATCHING

We propose a graph-based representation for the elliptic region shape descriptors introduced by Tuytelaars et al.13 In this representation we use image profiles to describe the relation between a pair of image regions. This new representation and a graph matching technique proposed in Ref. 1 are the basis of an object recognition method. An experimental comparative study between the original method and the new graph-based method is carried out. The results show that the graph-based method is more robust to scaling than the original method. Moreover, the misclassification rate using the graph-based method is considerably lower than that yielded by the original method.