Regions Of Discontinuity Research Articles

Geographic Patterns of Genetic Differentiation among Killer Whales in the Northern North Pacific

The difficulties associated with detecting population boundaries have long constrained the conservation and management of highly mobile, wide-ranging marine species, such as killer whales (Orcinus orca). In this study, we use data from 26 nuclear microsatellite loci and mitochondrial DNA sequences (988bp) to test a priori hypotheses about population subdivisions generated from a decade of killer whale surveys across the northern North Pacific. A total of 462 remote skin biopsies were collected from wild killer whales primarily between 2001 and 2010 from the northern Gulf of Alaska to the Sea of Okhotsk, representing both the piscivorous "resident" and the mammal-eating "transient" (or Bigg's) killer whales. Divergence of the 2 ecotypes was supported by both mtDNA and microsatellites. Geographic patterns of genetic differentiation were supported by significant regions of genetic discontinuity, providing evidence of population structuring within both ecotypes and corroborating direct observations of restricted movements of individual whales. In the Aleutian Islands (Alaska), subpopulations, or groups with significantly different mtDNA and microsatellite allele frequencies, were largely delimited by major oceanographic boundaries for resident killer whales. Although Amchitka Pass represented a major subdivision for transient killer whales between the central and western Aleutian Islands, several smaller subpopulations were evident throughout the eastern Aleutians and Bering Sea. Support for seasonally sympatric transient subpopulations around Unimak Island suggests isolating mechanisms other than geographic distance within this highly mobile top predator.

Partitioning the Genetic Diversity of a Virus Family: Approach and Evaluation through a Case Study of Picornaviruses

The recent advent of genome sequences as the only source available to classify many newly discovered viruses challenges the development of virus taxonomy by expert virologists who traditionally rely on extensive virus characterization. In this proof-of-principle study, we address this issue by presenting a computational approach (DEmARC) to classify viruses of a family into groups at hierarchical levels using a sole criterion-intervirus genetic divergence. To quantify genetic divergence, we used pairwise evolutionary distances (PEDs) estimated by maximum likelihood inference on a multiple alignment of family-wide conserved proteins. PEDs were calculated for all virus pairs, and the resulting distribution was modeled via a mixture of probability density functions. The model enables the quantitative inference of regions of distance discontinuity in the family-wide PED distribution, which define the levels of hierarchy. For each level, a limit on genetic divergence, below which two viruses join the same group, was objectively selected among a set of candidates by minimizing violations of intragroup PEDs to the limit. In a case study, we applied the procedure to hundreds of genome sequences of picornaviruses and extensively evaluated it by modulating four key parameters. It was found that the genetics-based classification largely tolerates variations in virus sampling and multiple alignment construction but is affected by the choice of protein and the measure of genetic divergence. In an accompanying paper (C. Lauber and A. E. Gorbalenya, J. Virol. 86:3905-3915, 2012), we analyze the substantial insight gained with the genetics-based classification approach by comparing it with the expert-based picornavirus taxonomy.

Eye movement patterns during the recognition of three-dimensional objects: Preferential fixation of concave surface curvature minima

This study used eye movement patterns to examine how high-level shape information is used during 3D object recognition. Eye movements were recorded while observers either actively memorized or passively viewed sets of novel objects, and then during a subsequent recognition memory task. Fixation data were contrasted against different algorithmically generated models of shape analysis based on: (1) regions of internal concave or (2) convex surface curvature discontinuity or (3) external bounding contour. The results showed a preference for fixation at regions of internal local features during both active memorization and passive viewing but also for regions of concave surface curvature during the recognition task. These findings provide new evidence supporting the special functional status of local concave discontinuities in recognition and show how studies of eye movement patterns can elucidate shape information processing in human vision.

DNA damage bypass operates in the S and G2 phases of the cell cycle and exhibits differential mutagenicity

Translesion DNA synthesis (TLS) employs low-fidelity DNA polymerases to bypass replication-blocking lesions, and being associated with chromosomal replication was presumed to occur in the S phase of the cell cycle. Using immunostaining with anti-replication protein A antibodies, we show that in UV-irradiated mammalian cells, chromosomal single-stranded gaps formed in S phase during replication persist into the G2 phase of the cell cycle, where their repair is completed depending on DNA polymerase ζ and Rev1. Analysis of TLS using a high-resolution gapped-plasmid assay system in cell populations enriched by centrifugal elutriation for specific cell cycle phases showed that TLS operates both in S and G2. Moreover, the mutagenic specificity of TLS in G2 was different from S, and in some cases overall mutation frequency was higher. These results suggest that TLS repair of single-stranded gaps caused by DNA lesions can lag behind chromosomal replication, is separable from it, and occurs both in the S and G2 phases of the cell cycle. Such a mechanism may function to maintain efficient replication, which can progress despite the presence of DNA lesions, with TLS lagging behind and patching regions of discontinuity.

Physalis method for heterogeneous mixtures of dielectrics and conductors: Accurately simulating one million particles using a PC

Prosperetti’s seminal Physalis method, an Immersed Boundary/spectral method, had been used extensively to investigate fluid flows with suspended solid particles. Its underlying idea of creating a cage and using a spectral general analytical solution around a discontinuity in a surrounding field as a computational mechanism to enable the accommodation of physical and geometric discontinuities is a general concept, and can be applied to other problems of importance to physics, mechanics, and chemistry. In this paper we provide a foundation for the application of this approach to the determination of the distribution of electric charge in heterogeneous mixtures of dielectrics and conductors. The proposed Physalis method is remarkably accurate and efficient. In the method, a spectral analytical solution is used to tackle the discontinuity and thus the discontinuous boundary conditions at the interface of two media are satisfied exactly. Owing to the hybrid finite difference and spectral schemes, the method is spectrally accurate if the modes are not sufficiently resolved, while higher than second-order accurate if the modes are sufficiently resolved, for the solved potential field. Because of the features of the analytical solutions, the derivative quantities of importance, such as electric field, charge distribution, and force, have the same order of accuracy as the solved potential field during postprocessing. This is an important advantage of the Physalis method over other numerical methods involving interpolation, differentiation, and integration during postprocessing, which may significantly degrade the accuracy of the derivative quantities of importance. The analytical solutions enable the user to use relatively few mesh points to accurately represent the regions of discontinuity. In addition, the spectral convergence and a linear relationship between the cost of computer memory/computation and particle numbers results in a very efficient method. In the present paper, the accuracy of the method is numerically investigated by example computations using one dielectric particle, one isolated conductor particle, one conductor particle connected to an external source with imposed voltage, and two conductor/dielectric particles with strong interactions. The efficiency of the method is demonstrated with one million particles, which suggests that the method can be used for many important engineering applications of broad interest.

Validation of the Swiss Monte Carlo Plan for a static and dynamic 6 MV photon beam

Monte Carlo (MC) based dose calculations can compute dose distributions with an accuracy surpassing that of conventional algorithms used in radiotherapy, especially in regions of tissue inhomogeneities and surface discontinuities. The Swiss Monte Carlo Plan (SMCP) is a GUI-based framework for photon MC treatment planning (MCTP) interfaced to the Eclipse treatment planning system (TPS). As for any dose calculation algorithm, also the MCTP needs to be commissioned and validated before using the algorithm for clinical cases. Aim of this study is the investigation of a 6 MV beam for clinical situations within the framework of the SMCP. In this respect, all parts i.e. open fields and all the clinically available beam modifiers have to be configured so that the calculated dose distributions match the corresponding measurements. Dose distributions for the 6 MV beam were simulated in a water phantom using a phase space source above the beam modifiers. The VMC++ code was used for the radiation transport through the beam modifiers (jaws, wedges, block and multileaf collimator (MLC)) as well as for the calculation of the dose distributions within the phantom. The voxel size of the dose distributions was 2 mm in all directions. The statistical uncertainty of the calculated dose distributions was below 0.4%. Simulated depth dose curves and dose profiles in terms of [Gy/MU] for static and dynamic fields were compared with the corresponding measurements using dose difference and γ analysis. For the dose difference criterion of ±1% of D max and the distance to agreement criterion of ±1 mm, the γ analysis showed an excellent agreement between measurements and simulations for all static open and MLC fields. The tuning of the density and the thickness for all hard wedges lead to an agreement with the corresponding measurements within 1% or 1 mm. Similar results have been achieved for the block. For the validation of the tuned hard wedges, a very good agreement between calculated and measured dose distributions was achieved using a 1%/1 mm criteria for the γ analysis. The calculated dose distributions of the enhanced dynamic wedges (10°, 15°, 20°, 25°, 30°, 45° and 60°) met the criteria of 1%/1 mm when compared with the measurements for all situations considered. For the IMRT fields all compared measured dose values agreed with the calculated dose values within a 2% dose difference or within 1 mm distance. The SMCP has been successfully validated for a static and dynamic 6 MV photon beam, thus resulting in accurate dose calculations suitable for applications in clinical cases.

Scanning SQUID Microscope NDT System Based on Eddy Current Testing

A scanning SQUID microscope based on HTS dc SQUID has been developed. The SQUID is mounted inside the insulation vacuum of a cryostat, which is separated from room temperature samples by a 65 μm thick sapphire window. Operating with a double-D exciting coil, it could be assembled to a scanning SQUID microscope Nondestructive Testing (NDT) system based on eddy current testing. The current is excited by an appropriate sinusoidal alternating (double-D coil) current in the conducting material, and then it induces a corresponding magnetic field. The vertical component of the field is then detected by the scanning SQUID system. The distortion of the field could be detected at the regions of discontinuities, such as flaws. Thereby, the defects inside the materials will be detected with neither contact nor destruction.

Inclined struts variation and its effect in reinforced concrete design

Cracks developing in high intensive shear areas are largely dictated by the effect of variable levels of flexural moments and their direction at the supports regions. Changes in developing crack inclinations, lead to changes in developing strut inclinations available for pressure, creating a sort of truss and allowing the transfer of forces for equilibrium. Strut and tie model analysis of reinforced concrete elements is considered an alternative to the usual approaches of analysis and design, and is applied effectively in regions of discontinuity. This analysis follows the truss analogy approach. Inclined concrete compression struts between parallel inclined cracks are assumed. The variation of the struts’ angle of inclination, depending on the shear and flexure levels, and its effect in reinforced concrete design are researched. This research introduces a rapid and approximated methodology for the representation of struts’ angle of inclination in analytical modelling of shear computations as customary in codes of practice

Numerical solutions for isoelectric focusing and isotachophoresis problems

This study combines an adaptive mesh redistribution (AMR) method and the space–time conservation element and solution element (CESE) method to construct a high-resolution scheme for the solution of electrophoresis pre-concentration and separation problems. In the proposed AMR–CESE scheme, the fine mesh points are moved toward the regions of discontinuity within the solution domain in accordance with the equidistribution principle. To reduce the numerical dissipation within the regions of the solution domain with a large spatial mesh, the spatial component of the CESE scheme is treated using a Courant–Friedrichs–Lewy (CFL) number insensitive scheme. The validity of the proposed approach is confirmed by comparing the results obtained for typical isoelectric focusing (IEF) and isotachophoresis (ITP) problems with those obtained from the conventional CESE scheme and the finite volume method (FVM), respectively. It is shown that the AMR–CESE scheme yields a better accuracy than uniform fixed-mesh solvers with no more than a minor increase in the computational cost.

High-resolution spectro-polarimetry of a flaring sunspot penumbra

We present simultaneous photospheric and chromospheric observations of the trailing sunspot in NOAA 10904 during a weak flare eruption (GOES magnitude B7.8), obtained with the Swedish Solar Telescope (SST) in La Palma, Canary Islands. High-resolution <i>H<i/> images show a typical two-ribbon structure that has been hitherto only known for larger flares, and the flare appears in a confined region that is discernible by a bright border. The underlying photosphere shows a disturbed penumbral structure with intersecting branches of penumbral filaments. High-resolution Doppler- and vector-magnetograms exhibit oppositely directed Evershed flows and magnetic field vectors in the individual penumbral branches, resulting in several regions of magnetic azimuth discontinuity and several islands where the vertical magnetic field is reversed. The discontinuity regions are co-spatial with the locations of the onset of the flare ribbons. From the results, we conclude that the confined flare region is detached from the global magnetic field structure by a separatrix marked by the bright border visible in <i>H<i/>. We further conclude that the islands of reversed vertical field appear because of flux emergence and that the strong magnetic shear appearing in the regions of magnetic azimuth discontinuity triggers the flare.

Reducing ghosting due to k-space discontinuities in fast spin echo (FSE) imaging by a new combination of k-space ordering and parallel imaging

In multi-echo imaging sequences like fast spin echo (FSE), the point spread function (PSF) in the phase encoding direction contains significant secondary peaks (sidebands). This is due to discontinuities in adjacent k-space data obtained at different echo times caused by T2 decay, and leads to ghosting and hence reduced image quality. Recently, utilising multiple coils for signal reception has become the standard configuration for MR systems due to the additional flexibility that parallel imaging (PI) methods can provide. PI methods generally obtain more data than is required to reconstruct an image. Here, this redundancy in information is exploited to reduce discontinuity-related ghosting in FSE imaging. Adjacent phase encoded k-space lines are acquired at different echo times alternately in the regions of discontinuity (called ‘feathering’). This moves the resulting ghost artefacts to the edges of the field of view. This property of the ghost then makes them amenable to removal using PI methods. With ‘feathered’ array coil data it is possible to reconstruct data over the region of the discontinuity from both echo times. By combining this data, a significant reduction in ghosting can be achieved. We show this approach to be effective through simulated and acquired MRI data.

Genome sequence of Desulfobacterium autotrophicum HRM2, a marine sulfate reducer oxidizing organic carbon completely to carbon dioxide.

Sulfate-reducing bacteria (SRB) belonging to the metabolically versatile Desulfobacteriaceae are abundant in marine sediments and contribute to the global carbon cycle by complete oxidation of organic compounds. Desulfobacterium autotrophicum HRM2 is the first member of this ecophysiologically important group with a now available genome sequence. With 5.6 megabasepairs (Mbp) the genome of Db. autotrophicum HRM2 is about 2 Mbp larger than the sequenced genomes of other sulfate reducers (SRB). A high number of genome plasticity elements (> 100 transposon-related genes), several regions of GC discontinuity and a high number of repetitive elements (132 paralogous genes Mbp−1) point to a different genome evolution when comparing with Desulfovibrio spp. The metabolic versatility of Db. autotrophicum HRM2 is reflected in the presence of genes for the degradation of a variety of organic compounds including long-chain fatty acids and for the Wood–Ljungdahl pathway, which enables the organism to completely oxidize acetyl-CoA to CO2 but also to grow chemolithoautotrophically. The presence of more than 250 proteins of the sensory/regulatory protein families should enable Db. autotrophicum HRM2 to efficiently adapt to changing environmental conditions. Genes encoding periplasmic or cytoplasmic hydrogenases and formate dehydrogenases have been detected as well as genes for the transmembrane TpII-c3, Hme and Rnf complexes. Genes for subunits A, B, C and D as well as for the proposed novel subunits L and F of the heterodisulfide reductases are present. This enzyme is involved in energy conservation in methanoarchaea and it is speculated that it exhibits a similar function in the process of dissimilatory sulfate reduction in Db. autotrophicum HRM2.

Towards a compact high-order method for non-linear hyperbolic systems. I: The Hermite Least-Square Monotone (HLSM) reconstruction

A new Hermite Least-Square Monotone (HLSM) reconstruction to calculate accurately complex flows on non-uniform meshes is presented.The coefficients defining the Hermite polynomial are calculated by using a least-square method. To introduce monotonicity conditions into the procedure, two constraints are added into the least-square system. Those constraints are derived by locally matching the high-order Hermite polynomial with a low-order TVD or ENO polynomial. To emulate these constraints only in regions of discontinuities, data-depending weights are defined; those weights are based upon normalized indicators of smoothness of the solution and are parameterized by a O(1) quantity. The reconstruction so generated is highly compact and is fifth-order accurate when the solution is smooth; this reconstruction becomes first-order in regions of discontinuities.By inserting this reconstruction into an explicit finite-volume framework, a spatially fifth-order non-oscillatory method is then generated. This method evolves in time the solution and its first derivative. In a one-dimensional context, a linear spectral analysis and extensive numerical experiments make it possible to assess the robustness and the advantages of the method in computing multi-scales problems with embedded discontinuities.

A compact high‐order HLL solver for nonlinear hyperbolic systems

AbstractWe present a new finite‐volume method for calculating complex flows on non‐uniform meshes. This method is designed to be highly compact and to accurately capture all discontinuities that may arise within the solution of a nonlinear hyperbolic system.In the first step, we devise a fourth‐degree Hermite polynomial to interpolate the solution. The coefficients defining this polynomial are calculated by using a least‐square method. To introduce monotonicity conditions within the procedure, two constraints are added into the least‐square system. Those constraints are derived by locally matching the high‐order Hermite polynomial with a low‐order TVD polynomial. To emulate these constraints only in regions of discontinuities, data‐depending weights are defined; these weights are based upon normalized indicators of smoothness of the solution and are parameterized by an O(1) quantity. The reconstruction so generated is highly compact and is fifth‐order accurate when the solution is smooth; this reconstruction becomes first order in regions of discontinuities.In the second step, this reconstruction is inserted in an HLL approximate Riemann solver. This solver is designed to correctly capture all discontinuities that may arise into the solution. To this aim, we introduce the contribution of a possible contact discontinuity into the HLL Riemann solver.Thus, a spatially fifth‐order non‐oscillatory method is generated. This method evolves in time the solution and its first derivative. In a one‐dimensional context, a linear spectral analysis and extensive numerical experiments make it possible to assess the robustness and the advantages of the method in computing multi‐scale problems with embedded discontinuities. Copyright © 2008 John Wiley &amp; Sons, Ltd.

Towards the automated analysis of regional aeromagnetic data to identify regions prospective for gold deposits

Quantitative analysis of geoscientific data to determine areas most likely to contain mineral deposits is becoming increasingly common in the mining industry. The approach is based on characterising areas known to contain deposits and seeking similar areas elsewhere. This paper presents an automatic image processing technique for the prospectivity analysis of Archaean lode-gold deposits, which differs from previous methods in that it is based solely on aeromagnetic data and does not require knowledge of the location of existing deposits. Instead, the aeromagnetic expressions of what are perceived to be geologically significant characteristics are sought within the aeromagnetic data. Gold mineralisation is known to occur near major crustal breaks manifesting as large-scale shear zones, which act as conduits for mineralising fluids. Mineralisation occurs in regions of structural complexity adjacent to the shear zones. Progressing towards the automatic detection of such regions, the proposed system finds firstly regions of magnetic discontinuity that correspond to both lithological boundaries and shear zones using a combination of texture analysis and symmetry feature detection techniques. Secondly, it examines the data using fractal analysis to find areas nearby with a complex magnetic expression (zones of structural complexity). The most prospective areas are those where inferred structural complexity occurs adjacent to the regions of magnetic discontinuity. A preliminary experiment was conducted using aeromagnetic data from the Yilgarn Craton in Western Australia and the regions selected by the proposed system contained over 76% of all known deposit locations and 82% of the greater than 1 tonne deposit locations.

Automatic Analysis of Aeromagnetic Images for Gold Exploration

Archaean lode gold mineralisation is known to occur near major crustal breaks manifesting as large-scale shear zones; which act as conduits for mineralising fluids. Mineralisation occurs in regions of structural complexity adjacent to the shear zones. This paper reports on-going image processing research that aims to automatically detect such geologically significant characteristics from aeromagnetic data for the prospectivity analysis. The current system finds the regions of magnetic discontinuity that are related to shear zones and lithological boundaries, and then identifies prospective regions nearby. The regions of magnetic discontinuity are detected using a combination of texture analysis and bilateral symmetric feature detection, where line-like features representing high local magnetic variations are identified. Then the areas adjacent to the regions of magnetic discontinuity are examined using a 2D fractal surface analysis technique to search for areas with complex magentic expressions that are related to structural complexity. A preliminary experiment was conducted using aeromagnetic data from the Yilgarn Craton in Western Australia and the regions selected by the proposed system contained over 76% of all known deposit locations and 82% of the greater than 1 tonne-deposit locations. Our on-going research also includes identifying the major shear zones from the regions of discontinuity to examine their spatial and geometrical characteristics.

Control of numerical effects of dispersion and dissipation in numerical schemes for efficient shock-capturing through an optimal Courant number

Composite schemes consist of several steps of a dispersive scheme followed by one step of a dissipative scheme [Liska Richard, Wendroff Burton. Composite schemes for conservation laws. SIAM J Numer Anal 1998;35(6):2250–71]. The latter [Liska Richard, Wendroff Burton. 2D shallow water equations by composite schemes. Int J Numer Meth Fluids 1999;30:461–79] acts as a filter reducing oscillations in regions of discontinuity. Liska and Wendroff have derived the composite Lax-Wendroff/Lax-Friedrichs (LWLF) [Liska Richard, Wendroff Burton, 1998] scheme which blends the Lax-Wendroff (LW) scheme with the 2-step Lax-Friedrichs (LF) scheme. The formulation of the 2-step Lax-Friedrichs scheme [Liska Richard, Wendroff Burton, 1998] is different from that of the classic Lax-Friedrichs scheme and has been devised by Liska [Liska Richard, Wendroff Burton, 1998]. In this work, we propose to replace LW scheme by MacCormack (MC) scheme since the latter is less dispersive. We obtain a new composite scheme in 1-D and in 2-D by blending the MacCormack scheme with the 2-step Lax-Friedrichs scheme which we term as the composite MacCormack/Lax-Friedrichs (MCLF) scheme. This is followed by analytical work on the effective amplification factor (EAF) and the relative phase error (RPE) for both families of schemes in 1-D and 2-D: LWLFn and MCLFn, consisting of ( n − 1) steps of the dispersive scheme (LW or MC) and 1 step of the dissipative LF scheme. We introduce a new concept, baptised as Curbing of Dispersion by Dissipation for Efficient Shock-capturing, CDDES in which a cfl number is computed whereby dissipation curbs dispersion. This cfl number is termed as optimal in this work. We conduct a comparative study based on numerical experiments in 2-D namely: contact-discontinuity problem [Ould Kaber SM. A legendre pseudospectral viscosity method. J Comput Phys 1996;128:165–80], rotating hill problem [Ould Kaber SM, 1996] and the deformative flow of Smolarkiewicz [Dabdub Donald, Seinfeld John H. Numerical advective schemes used in air quality models-sequential and parallel implementation. Atmos Environ 1994;28(20):3369–85, Ghods A, Sobouti F, Arkani-Hamed J. An improved second order method for solution of pure advection problems. Int J Numer Meth Fluids 2000;32:959–77, Nguyen K, Dabdub D. Two-level time-marching scheme using splines for solving the advection equation. Atmos Environ 2001;35:1627–37] to show that the MacCormack/Lax-Friedrichs (MCLF) scheme is more efficient than LWLF scheme to capture shocks in regions of discontinuity. We also show that better results are obtained at optimal cfl numbers for some variants of LWLFn and MCLFn schemes, with n = 2, 3, 4 and 5.

Contextual modulation of the motion aftereffect.

The authors examined center-surround effects for motion perception in human observers. The magnitude of the motion aftereffect (MAE) elicited by a drifting grating was measured with a nulling task and with a threshold elevation procedure. A surround grating of the same spatial frequency, temporal frequency, and orientation significantly reduced the magnitude of the MAE elicited by adaptation to the center grating. This effect was bandpass tuned for spatial frequency, orientation, and temporal frequency. Plaid surrounds but not contrast-modulated surrounds that moved in the same direction also reduced the MAE. These results provide psychophysical evidence for center-surround interactions analogous to those previously observed in electrophysiological studies of motion processing in primates. Collectively, these results suggest that motion processing, similar to texture processing, is organized for the purpose of highlighting regions of directional discontinuity in retinal images.

Effects of language experience and stimulus complexity on the categorical perception of pitch direction

Whether or not categorical perception results from the operation of a special, language-specific, speech mode remains controversial. In this cross-language (Mandarin Chinese, English) study of the categorical nature of tone perception, we compared native Mandarin and English speakers' perception of a physical continuum of fundamental frequency contours ranging from a level to rising tone in both Mandarin speech and a homologous (nonspeech) harmonic tone. This design permits us to evaluate the effect of language experience by comparing Chinese and English groups; to determine whether categorical perception is speech-specific or domain-general by comparing speech to nonspeech stimuli for both groups; and to examine whether categorical perception involves a separate categorical process, distinct from regions of sensory discontinuity, by comparing speech to nonspeech stimuli for English listeners. Results show evidence of strong categorical perception of speech stimuli for Chinese but not English listeners. Categorical perception of nonspeech stimuli was comparable to that for speech stimuli for Chinese but weaker for English listeners, and perception of nonspeech stimuli was more categorical for English listeners than was perception of speech stimuli. These findings lead us to adopt a memory-based, multistore model of perception in which categorization is domain-general but influenced by long-term categorical representations.

Generic object recognition with boosting

This paper explores the power and the limitations of weakly supervised categorization. We present a complete framework that starts with the extraction of various local regions of either discontinuity or homogeneity. A variety of local descriptors can be applied to form a set of feature vectors for each local region. Boosting is used to learn a subset of such feature vectors (weak hypotheses) and to combine them into one final hypothesis for each visual category. This combination of individual extractors and descriptors leads to recognition rates that are superior to other approaches which use only one specific extractor/descriptor setting. To explore the limitation of our system, we had to set up new, highly complex image databases that show the objects of interest at varying scales and poses, in cluttered background, and under considerable occlusion. We obtain classification results up to 81 percent ROC-equal error rate on the most complex of our databases. Our approach outperforms all comparable solutions on common databases.