Foreground Model Research Articles

Being Indian in post-colonial Hong Kong: Models of ethnicity, culture and language among Sindhis and Sikhs in Hong Kong

Using data from four focus groups, this paper examines the ways in which Hong Kong Sindhis and Sikhs talk about ethnic identities and their relationships to culture and language. It finds that in all the groups, a range of different ‘cultural models’ (Gee, 1999) of ethnicity, culture and language are drawn upon, with the same participants sometimes using several different models as they position themselves within the group discussions. However, the data also suggest that there may be a tendency for the two communities to foreground different models of ethnicity. Specifically, Sikh participants tend to draw more on an essentialist model of ethnicity than do the Sindhis, while the Sindhis seem to be more prepared to draw on more flexible models of the relationships between identity, culture and language.

Iterative Local-Global Energy Minimization for Automatic Extraction of Objects of Interest

We propose a novel global-local variational energy to automatically extract objects of interest from images. Previous formulations only incorporate local region potentials, which are sensitive to incorrectly classified pixels during iteration. We introduce a global likelihood potential to achieve better estimation of the foreground and background models and, thus, better extraction results. Extensive experiments demonstrate its efficacy.

Detectability of tensor modes in the presence of foregrounds

In inflationary models gravitational waves are produced in the early universe and generate $B$-type polarization in the cosmic microwave background (CMB). Since $B$ polarization is only generated by gravity waves it does not suffer from the usual cosmic variance. A perfect decomposition of the CMB into $B$-modes and $E$-modes would require data from the entire sky, which in practice is not possible because of the foreground contaminants. This leads to mixing of $E$ polarization into $B$, which introduces cosmic variance contamination of $B$ polarization and reduces sensitivity to gravity wave amplitude even in absence of detector noise. We present numerical results for the uncertainty in the tensor-to-scalar ratio using the Fisher matrix formalism for various resolutions and considering several cuts of the sky, using the foreground model based on dust maps and assuming 90 GHz operating frequency. We find that the usual scaling $△(\frac{T}{S})\ensuremath{\propto}{f}_{\mathrm{sky}}^{\ensuremath{-}1/2}$ is significantly degraded and becomes $△(\frac{T}{S})\ensuremath{\propto}{f}_{\mathrm{sky}}^{\ensuremath{-}2}$ for ${f}_{\mathrm{sky}}>0.7$. This dependence is affected only weakly by the choice of sky cuts. To put this into a context of what is required level of foreground cleaning, to achieve a $T/S={10}^{\ensuremath{-}3}$ detection at $3\ensuremath{\sigma}$ one needs to observe 15% of the sky as opposed to naive expectation of 0.3%. To prevent contamination over this large sky area at required level one must be able to remove polarized dust emission at or better than 0.1% of unpolarized intensity, assuming the cleanest part of the sky has been chosen. To achieve $T/S={10}^{\ensuremath{-}4}$ detection at $3\ensuremath{\sigma}$ one needs to observe 70% of the sky, which is only possible if dust emission is removed everywhere over this region at 0.01% level. Reaching $T/S={10}^{\ensuremath{-}2}$ should be easier: 1% of the sky is needed over which polarized emission needs to be removed at 1% of unpolarized intensity if the cleanest region is chosen. These results suggest that foreground contamination may make it difficult to achieve levels below $T/S={10}^{\ensuremath{-}3}$.

Bayesian modeling of dynamic scenes for object detection

Accurate detection of moving objects is an important precursor to stable tracking or recognition. In this paper, we present an object detection scheme that has three innovations over existing approaches. First, the model of the intensities of image pixels as independent random variables is challenged and it is asserted that useful correlation exists in intensities of spatially proximal pixels. This correlation is exploited to sustain high levels of detection accuracy in the presence of dynamic backgrounds. By using a nonparametric density estimation method over a joint domain-range representation of image pixels, multimodal spatial uncertainties and complex dependencies between the domain (location) and range (color) are directly modeled. We propose a model of the background as a single probability density. Second, temporal persistence is proposed as a detection criterion. Unlike previous approaches to object detection which detect objects by building adaptive models of the background, the foreground is modeled to augment the detection of objects (without explicit tracking) since objects detected in the preceding frame contain substantial evidence for detection in the current frame. Finally, the background and foreground models are used competitively in a MAP-MRF decision framework, stressing spatial context as a condition of detecting interesting objects and the posterior function is maximized efficiently by finding the minimum cut of a capacitated graph. Experimental validation of the proposed method is performed and presented on a diverse set of dynamic scenes.

NONPHOTOREALISTIC VIRTUAL ENVIRONMENT NAVIGATION FROM IMAGES

This paper presents a novel scheme based on a hybrid NPR/IBMR approach, for providing non-photorealistic virtual environment navigation from images. As a preprocess, we first divide the background and the foreground components in the image and then construct a simple 3D scene model consisting of background model and foreground object models, together with the corresponding non-photorealistic texture maps. Thus, through interactive navigation at runtime, non-photorealistic walk-through images for novel viewpoints can be generated in real time. To reduce visual artifacts in the resulting animation, we propose an effective method for generating multiresolution and view-dependent non-photorealistic texture maps that preserve the stroke size and coherence across different viewing resolutions and angles.

Identification of transcription factor binding sites with variable-order Bayesian networks

We propose a new class of variable-order Bayesian network (VOBN) models for the identification of transcription factor binding sites (TFBSs). The proposed models generalize the widely used position weight matrix (PWM) models, Markov models and Bayesian network models. In contrast to these models, where for each position a fixed subset of the remaining positions is used to model dependencies, in VOBN models, these subsets may vary based on the specific nucleotides observed, which are called the context. This flexibility turns out to be of advantage for the classification and analysis of TFBSs, as statistical dependencies between nucleotides in different TFBS positions (not necessarily adjacent) may be taken into account efficiently--in a position-specific and context-specific manner. We apply the VOBN model to a set of 238 experimentally verified sigma-70 binding sites in Escherichia coli. We find that the VOBN model can distinguish these 238 sites from a set of 472 intergenic 'non-promoter' sequences with a higher accuracy than fixed-order Markov models or Bayesian trees. We use a replicated stratified-holdout experiment having a fixed true-negative rate of 99.9%. We find that for a foreground inhomogeneous VOBN model of order 1 and a background homogeneous variable-order Markov (VOM) model of order 5, the obtained mean true-positive (TP) rate is 47.56%. In comparison, the best TP rate for the conventional models is 44.39%, obtained from a foreground PWM model and a background 2nd-order Markov model. As the standard deviation of the estimated TP rate is approximately 0.01%, this improvement is highly significant.

Low-order multipole maps of cosmic microwave background anisotropy derived fromWMAP

Recent analyses of the Wilkinson Microwave Anisotropy Probe (WMAP) data have suggested that the low-order multipoles of the cosmic microwave background (CMB) anisotropy distribution show cosmologically interesting and unexpected morphologies and amplitudes. In this paper, we apply a power equalization (PE) filter to the high-latitude WMAP data in order to reconstruct these low-l multipoles free from the largest Galactic foreground modelling uncertainties in the Galactic plane. The characteristic spatial distributions of the modes of order l = 2, 3, 4, 5 have been determined as a function of frequency, sky coverage and two methods for foreground correction: using the template-based corrections of Bennett et al. and by a simple linear projection scheme assuming the spectral dependence of the foreground components. Although the derived multipole maps are statistically consistent with previous estimates from Tegmark, de Oliveira-Costa & Hamilton (TOH) and Efstathiou, our analyses suggest that the K and Ka frequency bands remain significantly contaminated by residual foreground emission for the WMAP Kp2 mask. However, the l = 3, 4, 5 multipole maps for the Q, V and W channels indicate that, after foreground cleaning, these multipoles are dominated by the CMB anisotropy component. We confirm the TOH result that the octopole does indeed show structure in which its hot and cold spots are centred on a single plane in the sky, and show further that this is very stable with respect to the applied mask and foreground correction. The estimated quadrupole is much less stable showing non-negligible dependence on the Galactic foreground correction. Including these uncertainties is likely to weaken the statistical significance of the claimed alignment between the quadrupole and octopole. Nevertheless, these anisotropy patterns are also present in the COBE-DMR data and are unlikely to be associated with instrumental systematic artefacts.

Non-Gaussian signatures in the temperature fluctuation observed by the Wilkinson Microwave Anisotropy Probe

We present results from a test for the Gaussianity of the whole sky subdegree-scale cosmic microwave background (CMB) temperature anisotropy measured by the Wilkinson Microwave Anisotropy Probe (WMAP). We calculate the genus from the foreground-subtracted and Kp0-masked WMAP maps and measure the genus shift parameters defined at negative and positive threshold levels (Δν− and Δν+) and the asymmetry parameter (Δg) to quantify the deviation from the Gaussian relation. At WMAP Q, V and W bands, the genus and genus-related statistics imply that the observed CMB sky is consistent with a Gaussian random phase field. However, from the genus measurement on the Galactic northern and southern hemispheres, we have found two non-Gaussian signatures at the W-band resolution ( scale), i.e. the large difference of genus amplitudes between the north and the south and the positive genus asymmetry in the south, which are statistically significant at 2.6σ and 2.4σ levels, respectively. The large genus amplitude difference also appears in the WMAP Q- and V-band maps, deviating from the Gaussian prediction with a significance level of about 2σ. The probability that the genus curves show such a large genus amplitude difference exceeding the observed values at all Q, V and W bands in a Gaussian sky is only 1.4 per cent. Such non-Gaussian features are reduced as the higher Galactic cut is applied, but their dependence on the Galactic cut is weak. We discuss possible sources that can induce non-Gaussian features, such as the Galactic foregrounds, the integrated Sachs–Wolfe and the Sunyaev–Zel'dovich effects, and the reionization-induced low ℓ-mode non-Gaussianity that are aligned along the Galactic plane. We conclude that CMB data with higher signal-to-noise ratio and an accurate foreground model are needed to understand the non-Gaussian signatures.

Tracking multiple vehicles using foreground, background and motion models

In this paper a vehicle tracking algorithm is presented based on the combination of a novel per-pixel (Gaussian Mixture Based) background model and a set of foreground models of object size, position, velocity, and colour distribution. Each pixel in the scene is ‘explained’ as either background, belonging to a foreground object, or as noise. A projective ground-plane transform is used within the foreground model to strengthen object size and velocity consistency assumptions. A learned model of typical road travel direction and speed is used to provide a prior estimate of object velocity, which is used to initialise the velocity model for each of the foreground objects. The system runs at near video framerate (>20 fps) on modest hardware and is robust assuming sufficient image resolution is available and vehicle sizes do not greatly exceed the priors on object size used in object initialisation.

EGRET Upper Limits on the High‐Energy Gamma‐Ray Emission of Galaxy Clusters

We report EGRET upper limits on the high-energy gamma-ray emission from clusters of galaxies. EGRET observations between 1991 and 2000 were analyzed at positions of 58 individual clusters from a flux-limited sample of nearby X-ray bright galaxy clusters. Subsequently, a coadded image from individual galaxy clusters has been analyzed using an adequately adapted diffuse gamma-ray foreground model. The resulting 2 sigma upper limit for the average cluster is \~ 6 x 10^{-9} cm^{-2} s^{-1} for E > 100 MeV. Implications of the non--detection of prominent individual clusters and of the general inability to detect the X-ray brightest galaxy clusters as a class of gamma-ray emitters are discussed. We compare our results with model predictions on the high-energy gamma-ray emission from galaxy clusters as well as with recent claims of an association between unidentified or unresolved gamma-ray sources and Abell clusters of galaxies and find these contradictory.

Ultraviolet Spectra of Star‐forming Galaxies with Time‐dependent Dust Obscuration

We study the effect of time-dependent dust obscuration on synthetic ultraviolet line profiles of a young stellar population. If the youngest and most massive stars are more obscured than the older, less massive stars, the C IV ?1550 and other stellar wind lines are significantly diluted with respect to a simple foreground screen model for the dust. We propose to use stellar wind lines as a probe of the dust-obscuration model instead of the previously employed nebular emission lines. Since purely stellar diagnostics are utilized, uncertain assumptions on the nebular properties are unnecessary. Photoionization models demonstrate that the C IV ?1550 emission is typically dominated by stellar winds and nebular contamination is negligible. A first comparison with a galaxy sample observed with the Hopkins Ultraviolet Telescope favors a dust geometry affecting ionizing and nonionizing stars equally. We point out the need for higher quality data for a more rigorous comparison. The Hubble Space Telescope is capable of obtaining such data in the future.

Effects of Foreground Contamination on the Cosmic Microwave Background Anisotropy Measured byMAP

We study the effects of diffuse Galactic, far-infrared extragalactic source, and radio point source emission on the cosmic microwave background (CMB) anisotropy data anticipated from the MAP experiment. We focus on the correlation function and genus statistics measured from mock MAP foreground-contaminated CMB anisotropy maps generated in a spatially-flat cosmological constant dominated cosmological model. Analyses of the simulated MAP data at 90 GHz (0.3 deg FWHM resolution smoothed) show that foreground effects on the correlation function are small compared with cosmic variance. However, the Galactic emission, even just from the region with |b| > 20 deg, significantly affects the topology of CMB anisotropy, causing a negative genus shift non-Gaussianity signal. Given the expected level of cosmic variance, this effect can be effectively reduced by subtracting existing Galactic foreground emission models from the observed data. IRAS and DIRBE far-infrared extragalactic sources have little effect on the CMB anisotropy. Radio point sources raise the amplitude of the correlation function considerably on scales below 0.5 deg. Removal of bright radio sources above a 5 \sigma detection limit effectively eliminates this effect. Radio sources also result in a positive genus curve asymmetry (significant at 2 \sigma) on 0.5 deg scales. Accurate radio point source data is essential for an unambiguous detection of CMB anisotropy non-Gaussianity on these scales. Non-Gaussianity of cosmological origin can be detected from the foreground-subtracted CMB anisotropy map at the 2 \sigma level if the measured genus shift parameter |\Delta\nu| >= 0.02 (0.04) or if the measured genus asymmetry parameter |\Delta g| >= 0.03 (0.08) on a 0.3 (1.0) deg FWHM scale.

Galactic Contamination in the QMAP Experiment

We quantify the level of foreground contamination in the QMAP Cosmic Microwave Background (CMB) data with two objectives: (a) measuring the level to which the QMAP power spectrum measurements need to be corrected for foregrounds and (b) using this data set to further refine current foreground models. We cross-correlate the QMAP data with a variety of foreground templates. The 30 GHz Ka-band data is found to be significantly correlated with the Haslam 408 MHz and Reich and Reich 1420 MHz synchrotron maps, but not with the Diffuse Infrared Background Experiment (DIRBE) 240, 140 and 100 micron maps or the Wisconsin H-Alpha Mapper (WHAM) survey. The 40 GHz Q-band has no significant template correlations. We discuss the constraints that this places on synchrotron, free-free and dust emission. We also reanalyze the foreground-cleaned Ka-band data and find that the two band power measurements are lowered by 2.3% and 1.3%, respectively.

Foregrounds and Forecasts for the Cosmic Microwave Background

One of the main challenges facing upcoming cosmic microwave background (CMB) experiments will be to distinguish the cosmological signal from foreground contamination. We present a comprehensive treatment of this problem and study how foregrounds degrade the accuracy with which the Boomerang, MAP, and Planck experiments can measure cosmological parameters. Our foreground model includes not only the normalization, frequency dependence, and scale dependence for each physical component, but also variations in frequency dependence across the sky. When estimating how accurately cosmo- logical parameters can be measured, we include the important complication that foreground model parameters (we use about 500) must be simultaneously measured from the data as well. Our results are quite encouraging: despite all these complications, precision measurements of most cosmological param- eters are degraded by less than a factor of 2 for our main foreground model and by less than a factor of 5 in our most pessimistic scenario. Parameters measured though large-angle polarization signals suUer more degradation: up to 5 in the main model and 25 in the pessimistic case. The foregrounds that are potentially most damaging and therefore most in need of further study are vibrating dust emission and point sources, especially those in the radio frequencies. It is well known that E and B polarization contain valuable information about reionization and gravity waves, respectively. However, the cross- correlation between polarized and unpolarized foregrounds also deserves further study, as we —nd that it carries the bulk of the polarization information about most other cosmological parameters. Subject headings: cosmic microwave backgrounddiUuse radiationmethods: numerical ¨ polarization

Comparing and Combining Cosmic Microwave Background Data Sets

One of the best ways of spotting previously undetected systematic errors in cosmic microwave background (CMB) experiments is to compare two independent observations of the same region. We derive a set of tools for comparing and combining CMB data sets, applicable also in the common case where the two have different resolution or beam shape and therefore do not measure the same signal. We present a consistency test that is better than a χ2 test at detecting systematic errors. We show how two maps of different angular resolution can be combined without smoothing the higher resolution down to the lower one, and we generalize this to arbitrary beam configurations. We also show how lossless foreground removal can be performed even for foreground models involving scale dependence, latitude dependence, and spectral index variations in combination.

Evidence for an Intervening Stellar Population Toward the Large Magellanic Cloud

We identify a vertical extension of the red clump stars in the color magnitude diagram (CMD) of a section of the Large Magellanic Cloud (LMC). After subtracting the principal red clump component, we find a peak in the residual stellar distribution that is ~ 0.9 mag brighter than the peak of the principal red clump distribution. We consider and reject the following possible explanations for this population: inhomogeneous reddening, Galactic disk stars, random blends of red clump stars, correlated blends of red clump stars (binaries), evolution of the red clump stars, and red clump stars from a younger LMC stellar population. Combinations of these effects cannot be ruled out as the origin of this stellar population. A natural interpretation of this new population is that it consists of red clump stars that are closer to us than those in the LMC. We find corroborating evidence for this interpretation in Holtzman et al.'s (1997) Hubble Space Telescope CMD of the LMC field stars. The derived distance and projected angular surface density of these stars relative to the LMC stars (<~ 5 to 7%) are consistent with (1) models that attribute the observed microlensing lensing optical depth (Alcock et al. 1997) to a distinct foreground stellar population (Zhao 1997) and (2) tidal models of the interaction between the LMC and the Milky Way (Lin, Jones, & Klemola 1995). We conclude that the standard assumption of a smoothly distributed halo population out to the LMC cannot be substantiated without at least a detailed understanding of red clump stellar evolution, binary fractions, binary mass ratios, the spatial correlation of stars within the LMC, possible variations in the stellar populations of satellite galaxies, and differential reddening - all of which are highly complex.

Statistical mosaics for tracking

A method of robust feature-detection is proposed for visual tracking with a pan-tilt head. Even with good foreground models, the tracking process is liable to be disrupted by strong features in the background. Previous researchers have shown that the disruption can be somewhat suppressed by the use of image-subtraction. Building on this idea, a more powerful statistical model of background intensity is proposed in which a Gaussian mixture distribution is fitted to each of the pixels on a ‘virtual’ image plane. A fitting algorithm of the ‘Expectation-Maximisation’ type proves to be particularly effective here. Practical tests with contour tracking show marked improvement over image subtraction methods. Since the burden of computation is off-line, the online tracking process can run in real-time, at video field-rate.

M82 - The saga continues

We present new near-infrared observations of the starburst galaxy M 82, including high-resolution J, H, and K imagery, 1-5 micron spectrophotometry and photometry, and 1.05-1.35 micron spectroscopy. These measurements are used to examine the Stellar population in the galaxy and conditions in the interstellar medium. We combine our observations with results from the literature to derive a new set of constraints for starburst models, which are presented in a companion paper. Our measurements of Paschen line fluxes in combination with the strengths of other recombination lines allow study of the reddening in M 82. We find that a uniform foreground screen model for the dust geometry cannot explain the relative line strengths. A variety of other models can fit the data reasonably well, such as a nonuniform foreground screen or a homogeneous mix of dust and gas. The derived properties of the galaxy are not strongly dependent on which of these models is selected; we derive an ionizing flux of ~10^54^ s^-1^ and an absolute magnitude of the starburst region of M_K_ &lt; -22.5. Depending on the reddening model, the visual extinction to the nucleus lies between A_V_ = 12 and A_V_ = 27; we show that the nucleus does not coincide with any feature that is bright at visual wavelengths. The stellar H_2_O index in the nuclear region is less than 0.06 and the CO index is ~0.21; contamination of the CO band by emission by hot dust must be small. We use several lines in the 1.3 micron region to derive a lower limit to the electron density of n_e_&gt; 10^4.5^ cm^-3^ in the [Fe II] emission region. The [Fe II] line fluxes are consistent with excitation by fast shocks incident on dense gas. We report the tentative detection of several lines of H_2_, which, if confirmed, implicate UV fluorescence as the excitation mechanism.

An Analysis of Methods for Deriving the Constituent Insolation Components From Multipyranometer Array Measurements

A multipyranometer array (MPA) is a static platform of solar sensors designed to “untangle” the beam, diffuse and reflected portions of the total solar radiation. Over the past four years, a series of weather stations located throughout Colorado has used MPA’s to record insolation. This paper compares the beam radiation derived from an MPA with that measured by a nearby tracking pyrheliometer. Three different diffuse sky models as well as uniform and specular foreground reflectance models are used in the analysis. The ability of each model to predict the actual beam radiation is tabulated. Finally, two novel methods for deriving the beam component without the use of sky models are presented, along with a comparison with the conventional models.

The Cosmic Infrared Background

AbstractCosmic infrared background radiation, expected to carry the radiative record of luminous objects since they first formed, has yet to be detected. The Diffuse Infrared Background Experiment (DIRBE) on the Cosmic Background Explorer (COBE) satellite is designed to search for this primeval background over the spectral range from 1 to 300 μm. Initial examination of data from this experiment show that foreground radiations from the solar system and Galaxy dominate the diffuse infrared sky brightness, with relative minima near 3.5 μm and in the submillimeter wavelength range. DIRBE measurements do not confirm some previous data from rockets and IRAS. Preliminary upper limits on the cosmic infrared background are generally above theoretical expectations for pregalactic and protogalactic sources, and substantially above estimated lower limits based upon observations of external galaxies. Careful foreground modeling is required to reduce these limits or identify an isotropic residual.