https://doi.org/10.21437/interspeech.2014-267

In this work, we address the detection of vehicles in a video stream obtained from a moving airborne platform. We propose a Bayesian framework for estimating dense optical flow over time that explicitly estimates a persistent model of background appearance. The approach assumes that the scene can be described by background and occlusion layers, estimated within an expectation-maximization framework. The mathematical formulation of the paper is an extension of the work in (H. Yalcin et al., 2005) where motion and appearance models for foreground and background layers are estimated simultaneously in a Bayesian framework.

Both parametric and implicit representations can be used to model 2D curves and 3D surfaces. Each has certain advantages compared to the other. Implicit polynomial (IP) methods are not as popular as parametric procedures because the lack of general procedures for obtaining IP models of higher degree has prevented their general use in many practical applications. In most cases today, parametric equations are used to model curves and surfaces. One such parametric representation, elliptic Fourier descriptors (EFD) has been widely used to represent 2D and 3D curves, as well as 3D surfaces. Although EFDs can represent nearly all curves, it is often convenient to have an implicit algebraic description, F(x,y)=0, especially for determining whether given points lie on the curve. Algebraic curves and surfaces have proven very useful also in many model-based applications. Various algebraic and geometric invariants obtained from these implicit models have been studied rather extensively. We present a new non-symbolic implicitization technique called the matrix annihilation method, for converting parametric Fourier representations to implicit polynomial form.

The ability to combine vision and behavior has proven to be vital for a system to achieve real-time performance on a variety of tasks. Using biological evidence, it has been shown that such systems must have oculomotion, attention and spatio-temporal reasoning capabilities. The paper describes BUVIS an automated visual inspection system-which is designed to have these capabilities-and selective perceptual processing thus achieved can result in real-time behavior in visual inspection of manufactured parts.