Bimodal biometric verification based on face and lips

In this paper, we present a novel approach to person verification by fusing face and lip features. Specifically, the face is modeled by the discriminative common vector and the discrete wavelet transform. Our lip features are simple geometric features based on a lip contour, which can be interpreted as multiple spatial widths and heights from a center of mass. In order to combine these features, we consider two simple fusion strategies: data fusion before training and score fusion after training, working with two different face databases. Fusing them together boosts the performance to achieve an equal error rate as low as 0.4% and 0.28%, respectively, confirming that our approach of fusing lips and face is effective and promising.

Robust bimodal person identification using face and speech with limited training data and corruption of both modalities

This paper presents a novel method of audio-visual fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there is a limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new representation and a modified cosine similarity are introduced for combining and comparing bimodal features with limited training data as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal data set created from the SPIDRE and AR databases with variable noise corruption of speech and occlusion in the face images. The new method has demonstrated improved recognition accuracy.

The VLT-FLAMES tarantula survey:X. Evidence for a bimodal distribution of rotational velocities for the single early B-type stars

Aims. Projected rotational velocities (ve sin i) have been estimated for 334 targets in the VLT-FLAMES Tarantula Survey that do not manifest significant radial velocity variations and are not supergiants. They have spectral types from approximately O9.5 to B3. The estimates have been analysed to infer the underlying rotational velocity distribution, which is critical for understanding the evolution of massive stars. Methods. Projected rotational velocities were deduced from the Fourier transforms of spectral lines, with upper limits also being obtained from profile fitting. For the narrower lined stars, metal and non-diffuse helium lines were adopted, and for the broader lined stars, both non-diffuse and diffuse helium lines; the estimates obtained using the different sets of lines are in good agreement. The uncertainty in the mean estimates is typically 4% for most targets. The iterative deconvolution procedure of Lucy has been used to deduce the probability density distribution of the rotational velocities. Results. Projected rotational velocities range up to approximately 450 kms-1 and show a bi-modal structure. This is also present in the inferred rotational velocity distribution with 25% of the sample having 0 <ve <100 km s^-1 and the high velocity component having v_e ∼ 250 km s^-1. There is no evidence from the spatial and radial velocity distributions of the two components that they represent either field and cluster populations or different episodes of star formation. Be-type stars have also been identified. Conclusions. The bi-modal rotational velocity distribution in our sample resembles that found for late-B and early-A type stars.While magnetic braking appears to be a possible mechanism for producing the low-velocity component, we can not rule out alternative explanations. © ESO 2013.

What motion distributions yield global transparency and spatial segmentation?

We have examined the ability of observers to parse bimodal local-motion distributions into two global motion surfaces, either overlapping (yielding transparent motion) or spatially segregated (yielding a motion boundary). The stimuli were random dot kinematograms in which the direction of motion of each dot was drawn from one of two rectangular probability distributions. A wide range of direction distribution widths and separations was tested. The ability to discriminate the direction of motion of one of the two motion surfaces from the direction of a comparison stimulus was used as an objective test of the perception of two discrete surfaces. Performance for both transparent and spatially segregated motion was remarkably good, being only slightly inferior to that achieved with a single global motion surface. Performance was consistently better for segregated motion than for transparency. Whereas transparent motion was only perceived with direction distributions which were separated by a significant gap, segregated motion could be seen with abutting or even partially overlapping direction distributions. For transparency, the critical gap increased with the range of directions in the distribution. This result does not support models in which transparency depends on detection of a minimum size of gap defining a bimodal direction distribution. We suggest, instead, that the operations which detect bimodality are scaled (in the direction domain) with the overall range of distributions. This yields a flexible, adaptive system that determines whether a gap in the direction distribution serves as a segmentation cue or is smoothed as part of a unitary computation of global motion.

Progressive loss of ferroelectricity under bipolar pulsed fields and experimental determination of non-switchable polarization in Au/Ba0.5Sr0.5TiO3/SrRuO3 thin-film capacitors

175 nm-thick Ba0.5Sr0.5TiO3 (BST) thin film fabricated by pulsed laser deposition (PLD) technique is found to be a mixture of two distributions of material. We discuss whether these two components are nano-regions of paraelectric and ferroelectric phases, or a bimodal grain-size distribution, or an effect of oxygen vacancy gradient from the electrode interface. The fraction of switchable ferroelectric phase decreases under bipolar pulsed fields, but it recovers after removal of the external fields. The plot of capacitance in decreasing dc voltage (C(Vdown arrow) versus that in increasing dc 61 voltage C(Vup arrow) is a superposition of overlapping of two triangles, in contrast to one well-defined triangle for typical ferroelectric SrBi2Ta2O9 thin films.

Half-precessional dynamics of monsoon rainfall near the East African Equator

External climate forcings-such as long-term changes in solar insolation-generate different climate responses in tropical and high latitude regions(1). Documenting the spatial and temporal variability of past climates is therefore critical for understanding how such forcings are translated into regional climate variability. In contrast to the data-richmiddle and high latitudes, high-quality climate-proxy records from equatorial regions are relatively few(2-4), especially from regions experiencing the bimodal seasonal rainfall distribution associated with twice-annual passage of the Intertropical Convergence Zone. Here we present a continuous and well-resolved climate-proxy record of hydrological variability during the past 25,000 years from equatorial East Africa. Our results, based on complementary evidence from seismic-reflection stratigraphy and organic biomarker molecules in the sediment record of Lake Challa near Mount Kilimanjaro, reveal that monsoon rainfall in this region varied at half-precessional (similar to 11,500-year) intervals in phase with orbitally controlled insolation forcing. The southeasterly and northeasterly monsoons that advect moisture from the western Indian Ocean were strengthened in alternation when the inter-hemispheric insolation gradient was at a maximum; dry conditions prevailed when neither monsoon was intensified and modest local March or September insolation weakened the rain season that followed. On sub-millennial timescales, the temporal pattern of hydrological change on the East African Equator bears clear high-northern-latitude signatures, but on the orbital timescale it mainly responded to low-latitude insolation forcing. Predominance of low-latitude climate processes in this monsoon region can be attributed to the low-latitude position of its continental regions of surface air flow convergence, and its relative isolation from the Atlantic Ocean, where prominent meridional overturning circulation more tightly couples low-latitude climate regimes to high-latitude boundary conditions.

Rapid sympatric ecological differentiation of crater lake cichlid fishes within historic times

Background: After a volcano erupts, a lake may form in the cooled crater and become an isolated aquatic ecosystem. This makes fishes in crater lakes informative for understanding sympatric evolution and ecological diversification in barren environments. From a geological and limnological perspective, such research offers insight about the process of crater lake ecosystem establishment and speciation. In the present study we use genetic and coalescence approaches to infer the colonization history of Midas cichlid fishes (Amphilophus cf. citrinellus) that inhabit a very young crater lake in Nicaragua-the ca. 1800 year-old Lake Apoyeque. This lake holds two sympatric, endemic morphs of Midas cichlid: one with large, hypertrophied lips (~20% of the total population) and another with thin lips. Here we test the associated ecological, morphological and genetic diversification of these two morphs and their potential to represent incipient speciation.
Results: Gene coalescence analyses [11 microsatellite loci and mitochondrial DNA (mtDNA) sequences] suggest that crater lake Apoyeque was colonized in a single event from the large neighbouring great lake Managua only about 100 years ago. This founding in historic times is also reflected in the extremely low nuclear and mitochondrial genetic diversity in Apoyeque. We found that sympatric adult thin- and thick-lipped fishes occupy distinct ecological trophic niches. Diet, body shape, head width, pharyngeal jaw size and shape and stable isotope values all differ significantly between the two lip-morphs. The eco-morphological features pharyngeal jaw shape, body shape, stomach contents and stable isotopes (d15N) all show a bimodal distribution of traits, which is compatible with the expectations of an initial stage of ecological speciation under disruptive selection. Genetic differentiation between the thin- and thick-lipped population is weak at mtDNA sequence (FST = 0.018) and absent at nuclear microsatellite loci (FST < 0.001).
Conclusions: This study provides empirical evidence of eco-morphological differentiation occurring very quickly after the colonization of a new and vacant habitat. Exceptionally low levels of neutral genetic diversity and inference from coalescence indicates that the Midas cichlid population in Apoyeque is much younger (ca. 100 years or generations old) than the crater itself (ca. 1 800 years old). This suggests either that the crater remained empty for many hundreds of years after its formation or that remnant volcanic activity prevented the establishment of a stable fish population during the early life of the crater lake. Based on our findings of eco-morphological variation in the Apoyeque Midas cichlids, and known patterns of adaptation in Midas cichlids in general, we suggest that this population may be in a very early stage of speciation (incipient species), promoted by disruptive selection and ecological diversification.

Combining multisensory temporal information for movement synchronisation

The ability to synchronise actions with environmental events is a fundamental skill supporting a variety of group activities. In such situations, multiple sensory cues are usually available for synchronisation, yet previous studies have suggested that auditory cues dominate those from other modalities. We examine the control of rhythmic action on the basis of auditory and haptic cues and show that performance is sensitive to both sources of information for synchronisation. Participants were required to tap the dominant hand index finger in synchrony with a metronome defined by periodic auditory tones, imposed movements of the non-dominant index finger, or both cues together. Synchronisation was least variable with the bimodal metronome as predicted by a maximum likelihood estimation (MLE) model. However, increases in timing variability of the auditory cue resulted in some departures from the MLE model. Our findings indicate the need for further investigation of the MLE account of the integration of multisensory signals in the temporal control of action.

Growth-phase regulation of lipopolysaccharide O-antigen chain length influences serum resistance in serovars of Salmonella

The amount of lipopolysaccharide (LPS) O antigen (OAg) and its chain length distribution are important factors that protect bacteria from serum complement. Salmonella enterica serovar Typhi produces LPS with long chain length distribution (L-OAg) controlled by the wzz gene, whereas serovar Typhimurium produces LPS with two OAg chain lengths: an L-OAg controlled by Wzz(ST) and a very long (VL) OAg determined by Wzz(fepE). This study shows that serovar Enteritidis also has a bimodal OAg distribution with two preferred OAg chain lengths similar to serovar Typhimurium. It was reported previously that OAg production by S. Typhi increases at the late exponential and stationary phases of growth. The results of this study demonstrate that increased amounts of L-OAg produced by S. Typhi grown to stationary phase confer higher levels of bacterial resistance to human serum. Production of OAg by serovars Typhimurium and Enteritidis was also under growth-phase-dependent regulation; however, while the total amount of OAg increased during growth, the VL-OAg distribution remained constant. The VL-OAg distribution was primarily responsible for complement resistance, protecting the non-typhoidal serovars from the lytic action of serum irrespective of the growth phase. As a result, the non-typhoidal species were significantly more resistant than S. Typhi to human serum. When S. Typhi was transformed with a multicopy plasmid containing the S. Typhimurium wzz(fepE) gene, resistance to serum increased to levels comparable to the non-typhoidal serovars. In contrast to the relevant role for high-molecular-mass OAg molecules, the presence of Vi antigen did not contribute to serum resistance of clinical isolates of serovar Typhi.

O-antigen modal chain length in Shigella flexneri 2a is growth-regulated through RfaH-mediated transcriptional control of the wzy gene

Shigella flexneri 2a 2457T produces lipopolysaccharide (LPS) with two O-antigen (OAg) chain lengths: a short (S-OAg) controlled by WzzB and a very long (VL-OAg) determined by Wzz(pHS-2). This study demonstrates that the synthesis and length distribution of the S. flexneri OAg are under growth-phase-dependent regulation. Quantitative electrophoretic analysis showed that the VL-OAg increased during growth while the S-OAg distribution remained constant. Increased production of VL-OAg correlated with the growth-phase-regulated expression of the transcription elongation factor RfaH, and was severely impaired in a DeltarfaH mutant, which synthesized only low-molecular-mass OAg molecules and a small amount of S-OAg. Real-time RT-PCR revealed a drastic reduction of wzy polymerase gene expression in the DeltarfaH mutant. Complementation of this mutant with the wzy gene cloned into a high-copy-number plasmid restored the bimodal OAg distribution, suggesting that cellular levels of Wzy influence not only OAg polymerization but also chain-length distribution. Accordingly, overexpression of wzy in the wild-type strain resulted in production of a large amount of high-molecular-mass OAg molecules. An increased dosage of either wzzB or wzz(pHS-2) also altered OAg chain-length distribution. Transcription of wzzB and wzz(pHS-2) genes was regulated during bacterial growth but in an RfaH-independent manner. Overall, these findings indicate that expression of the wzy, wzzB and wzz(pHS-2) genes is finely regulated to determine an appropriate balance between the proteins responsible for polymerization and chain-length distribution of S. flexneri OAg.

Effect of microstructure on the fatigue properties of Ti-6Al-4V titanium alloys

Through an analysis on microstructure and high cycle fatigue (HCF) properties of Ti-6Al-4V alloys which were selected from literature, the effects of microstructure types and microstructure parameters on HCF properties were investigated systematically. The results show that the HCF properties are strongly determined by microstructure types for Ti-6Al-4V. Generally the HCF strengths of different microstructures decrease in the order of bimodal, lamellar and equiaxed microstructure. Additionally, microstructure parameters such as the primary a (a) content and the a grain size in bimodal microstructures, the a lamellar width in lamellar microstructure and the a grain size in equiaxed microstructures, can influence the HCF properties. © 2012 Elsevier Ltd.

Robust Multimodal Person Identification With Limited Training Data

This paper presents a novel method of audio-visual feature-level fusion for person identification where both the speech and facial modalities may be corrupted, and there is a lack of prior knowledge about the corruption. Furthermore, we assume there are limited amount of training data for each modality (e.g., a short training speech segment and a single training facial image for each person). A new multimodal feature representation and a modified cosine similarity are introduced to combine and compare bimodal features with limited training data, as well as vastly differing data rates and feature sizes. Optimal feature selection and multicondition training are used to reduce the mismatch between training and testing, thereby making the system robust to unknown bimodal corruption. Experiments have been carried out on a bimodal dataset created from the SPIDRE speaker recognition database and AR face recognition database with variable noise corruption of speech and occlusion in the face images. The system's speaker identification performance on the SPIDRE database, and facial identification performance on the AR database, is comparable with the literature. Combining both modalities using the new method of multimodal fusion leads to significantly improved accuracy over the unimodal systems, even when both modalities have been corrupted. The new method also shows improved identification accuracy compared with the bimodal systems based on multicondition model training or missing-feature decoding alone.