this paper, may not be detailed enough for the synthesis of clutter images

Seeking probability models for images, we employ a spectral approach where the images are decomposed using bandpass filters and probability models are imposed on the filter outputs (also called spectral components). We employ a (two-parameter) family of probability densities, introduced in [11] and called Bessel K forms, for modeling the marginal densities of the spectral components, and demonstrate their fit to the observed histograms for video, infrared, and range images. Motivated by object-based models for image analysis, a relationship between the Bessel parameters and the imaged objects is established. Using 2-metric on the set of Bessel K forms, we propose a pseudometric on the image space for quantifying image similarities/differences. Some applications, including clutter classification and pruning of hypotheses for target recognition, are presented.

In the problem of recognizing targets from their observed images, the estimation of target orientations, as elements of the rotation group SO(3), plays an important role. For k- objects the unknown parameter is an element of SO(3) k . Since k may be unknown a-priori, the parameter space is extended to X = [ 1 k=0 SO(3) k . In this representation, both the target orientations and their numbers have to be estimated simultaneously. We present a Bayesian approach which builds a posterior probability measure on X . Then, utilizing a Markov jump-diffusion process X(t), we sample from this posterior to empirically generate the estimates. The two components of X(t), jumps and diffusions, are chosen in such a way that the resulting Markov process has the desired ergodic property: averages along its sample paths converge to the expectations under the posterior. Proper choice of the diffusion parameters and the jump intensities is demonstrated and the ergodic result associated wi...

A Bayesian approach to identifying faces from their IR facial images amounts to testing of discrete hypotheses in presence of nuisance variables such as pose, facial expression, and thermal state. We propose an efficient, lowlevel technique for hypothesis pruning, i.e. shortlisting high probability subjects, from given observed image(s). (This subset can be further tested using some detailed high-level model for eventual identification). Hypothesis pruning is accomplished using wavelet decompositions (of the observed images) followed by analysis of lower-order statistics of the coefficients. Specifically, we filter infrared (IR) images using bandpass filters and model the marginal densities of the outputs via a parametric family that was introduced in [11]. IR images are compared using an L²-metric computed directly from the parameters. Results from experiments on IR face identification and statistical pruning are presented.

this article. Mallat [9], Mumford et al. [10], Wainwright et al. [11], Field [12], [13], and others have extensively studied empirical distributions of images taken from large databases and have discovered certain interesting patterns. They have established that image statistics under common representations, such as wavelets or subspace bases (PCA, ICA, Fisher's etc.), point to non-Gaussian distributions. For example, a popular mechanism for decomposing images locally, in space and frequency, using wavelet transforms leads to coefficients that are quite non-Gaussian. The histograms display heavy tails and sharp cusps at the median. It is imperative that any probability model adopted for image analysis should explain such observed phenomena

Tracking of target pose is important for ATR in situations where there is a relative motion between the targets and the sensor#s#. Taking a Bayesian approach, we formulate the problem of jointly tracking the target positions and orientations #as elements of SE#3## as a problem in nonlinear #ltering. Combining pertinent ideas from importance sampling and sequential methods, we apply an iterative Monte Carlo approach to solve for MMSE solutions. This tracking algorithm is demonstrated for tracking individual targets in a simulated environment. Keywords: ATR, Target pose tracking, nonlinear #ltering, sequential Monte Carlo methods, Newtonian dynamics 1. INTRODUCTION In automated target recognition #ATR#, the goal is to analyze observed images and recognize targets of interest contained in them. In the process of target recognition, estimation of nuisance parameters, such as target position and orientation, plays an important role. It is well known that a more accurate parameter estimat...

This paper proposes a unique map learning method for mobile robots based on the co-visibility information of objects i.e., the information on whether two objects are visible at the same time or not from the current position. This method first estimates empirical distances among the objects using a simple heuristics -- "a pair of objects observed at the same time more frequently is likely to be located more closely together". Then it computes all the coordinates of the objects by multidimensional scaling (MDS) technique. In the latter part of this paper, it is shown that the proposed method is able to learn qualitatively very accurate maps though it uses only such primitive information, and that it is robust against some kinds of object recognition errors.

Investigators interested in model order estimation have tended to divide themselves into widely separated camps; this survey of the contributions of Schwarz, Wallace, Rissanen, and their coworkers attempts to build bridges between the various viewpoints, illuminating connections which may have previously gone unnoticed and clarifying misconceptions which seem to have propagated in the applied literature. Our tour begins with Schwarz's approximation of Bayesian integrals via Laplace's method. We then introduce the concepts underlying Rissanen 's minimum description length principle via a Bayesian scenario with a known prior; this provides the groundwork for understanding his more complex non-Bayesian MDL which employs a "universal" encoding of the integers. Rissanen's method of parameter truncation is contrasted with that employed in various versions of Wallace's minimum message length criteria.

This paper examines metrics for measuring clutter effectiveness on model-based automatic target recognition systems with FLIR sensors. The measure for clutter effectiveness proposed is the difference of two Kullback-Leibler distances between the idealized approximate probabilistic models without clutter and the real model containing clutter. We establish that occluding objects and clutter, when manipulated, do not present a fundamental challenge to model-based ATR system if the model manipulated is an accurate representation of the obscuring clutter. However, if the obscurer is not manipulated, performance degrades in cases where the obscurer is an "effective clutterer." To quantify the effect of clutter in ATR, estimation and detection problems are considered for rigid ground-based targets. For estimating the orientation of a vehicle, the Hilbert-Schmidt distance is employed. Aaron D. Lanterman, Univ. of Illinois, Coordinated Science Laboratory, 1308 W. Main, Urbana, IL 61801, E-mai...

this paper. 129 in encoding y using q(y) is \Gamma ln q(y) + ln p l (yjx(y)) = ln