Unconstrained and Contactless Hand Geometry Biometrics

This paper presents a hand biometric system for contact-less, platform-free scenarios, proposing innovative methods in feature extraction, template creation and template matching. The evaluation of the proposed method considers both the use of three contact-less publicly available hand databases, and the comparison of the performance to two competitive pattern recognition techniques existing in literature: namely Support Vector Machines (SVM) and k-Nearest Neighbour (k-NN). Results highlight the fact that the proposed method outcomes existing approaches in literature in terms of computational cost, accuracy in human identification, number of extracted features and number of samples for template creation. The proposed method is a suitable solution for human identification in contact-less scenarios based on hand biometrics, providing a feasible solution to devices with limited hardware requirements like mobile devices

Text categorization methods for automatic estimation of verbal intelligence

In this paper we investigate whether conventional text categorization methods may suffice to infer different verbal intelligence levels. This research goal relies on the hypothesis that the vocabulary that speakers make use of reflects their verbal intelligence levels. Automatic verbal intelligence estimation of users in a spoken language dialog system may be useful when defining an optimal dialog strategy by improving its adaptation capabilities. The work is based on a corpus containing descriptions (i.e. monologs) of a short film by test persons yielding different educational backgrounds and the verbal intelligence scores of the speakers. First, a one-way analysis of variance was performed to compare the monologs with the film transcription and to demonstrate that there are differences in the vocabulary used by the test persons yielding different verbal intelligence levels. Then, for the classification task, the monologs were represented as feature vectors using the classical TF–IDF weighting scheme. The Naive Bayes, k-nearest neighbors and Rocchio classifiers were tested. In this paper we describe and compare these classification approaches, define the optimal classification parameters and discuss the classification results obtained.

Investigating verbal intelligence using the TF-IDF approach

In this paper we investigated differences in language use of speakers yielding different verbal intelligence when they describe the same event. The work is based on a corpus containing descriptions of a short film and verbal intelligence scores of the speakers. For analyzing the monologues and the film transcript, the number of reused words, lemmas, n-grams, cosine similarity and other features were calculated and compared to each other for different verbal intelligence groups. The results showed that the similarity of monologues of higher verbal intelligence speakers was greater than of lower and average verbal intelligence participants. A possible explanation of this phenomenon is that candidates yielding higher verbal intelligence have a better short-term memory. In this paper we also checked a hypothesis that differences in vocabulary of speakers yielding different verbal intelligence are sufficient enough for good classification results. For proving this hypothesis, the Nearest Neighbor classifier was trained using TF-IDF vocabulary measures. The maximum achieved accuracy was 92.86%.

Incorporating Group Recommendations to Recommender Systems: Alternatives and Performance

La importancia de los sistemas de recomendación ha experimentado un crecimiento exponencial como consecuencia del auge de las redes sociales. En esta tesis doctoral presentaré una amplia visión sobre el estado del arte de los sistemas de recomendación. Incialmente, estos estaba basados en fitrado demográfico, basado en contendio o colaborativo. En la actualidad, estos sistemas incorporan alguna información social al proceso de recomendación. En el futuro utilizarán información implicita, local y personal proveniente del Internet de las cosas. Los sistemas de recomendación basados en filtrado colaborativo se pueden modificar con el fin de realizar recomendaciones a grupos de usuarios. Existen trabajos previos que han incluido estas modificaciones en diferentes etapas del algoritmo de filtrado colaborativo: búsqueda de los vecinos, predicción de las votaciones y elección de las recomendaciones. En esta tesis doctoral proporcionaré un nuevo método que realizar el proceso de unficación (pasar de varios usuarios a un grupo) en el primer paso del algoritmo de filtrado colaborativo: cálculo de la métrica de similaridad. Proporcionaré una formalización completa del método propuesto. Explicaré cómo obtener el conjunto de k vecinos del grupo de usuarios y mostraré cómo obtener recomendaciones usando dichos vecinos. Asimismo, incluiré un ejemplo detallando cada paso del método propuesto en un sistema de recomendación compuesto por 8 usuarios y 10 items. Las principales características del método propuesto son: (a) es más rápido (más eficiente) que las alternativas proporcionadas por otros autores, y (b) es al menos tan exacto y preciso como otras soluciones estudiadas. Para contrastar esta hipótesis realizaré varios experimentos que miden la precisión, la exactitud y el rendimiento del método. Los resultados obtenidos se compararán con los resultados de otras alternativas utilizadas en la recomendación de grupos. Los experimentos se realizarán con las bases de datos de MovieLens y Netflix. ABSTRACT The importance of recommender systems has grown exponentially with the advent of social networks. In this PhD thesis I will provide a wide vision about the state of the art of recommender systems. They were initially based on demographic, contentbased and collaborative filtering. Currently, these systems incorporate some social information to the recommendation process. In the future, they will use implicit, local and personal information from the Internet of Things. As we will see here, recommender systems based on collaborative filtering can be used to perform recommendations to group of users. Previous works have made this modification in different stages of the collaborative filtering algorithm: establishing the neighborhood, prediction phase and determination of recommended items. In this PhD thesis I will provide a new method that carry out the unification process (many users to one group) in the first stage of the collaborative filtering algorithm: similarity metric computation. I will provide a full formalization of the proposed method. I will explain how to obtain the k nearest neighbors of the group of users and I will show how to get recommendations using those users. I will also include a running example of a recommender system with 8 users and 10 items detailing all the steps of the method I will present. The main highlights of the proposed method are: (a) it will be faster (more efficient) that the alternatives provided by other authors, and (b) it will be at least as precise and accurate as other studied solutions. To check this hypothesis I will conduct several experiments measuring the accuracy, the precision and the performance of my method. I will compare these results with the results generated by other methods of group recommendation. The experiments will be carried out using MovieLens and Netflix datasets.

Maintaining distances with the engineer: patterns of coexistence in plant communities beyond the patch-bare dichotomy

Two-phase plant communities with an engineer conforming conspicuous patches and affecting the performance and patterns of coexisting species are the norm under stressful conditions. To unveil the mechanisms governing coexistence in these communities at multiple spatial scales, we have developed a new point-raster approach of spatial pattern analysis, which was applied to a Mediterranean high mountain grassland to show how Festuca curvifolia patches affect the local distribution of coexisting species. We recorded 22 111 individuals of 17 plant perennial species. Most coexisting species were negatively associated with F. curvifolia clumps. Nevertheless, bivariate nearest-neighbor analyses revealed that the majority of coexisting species were confined at relatively short distances from F. curvifolia borders (between 0-2 cm and up to 8 cm in some cases). Our study suggests the existence of a fine-scale effect of F. curvifolia for most species promoting coexistence through a mechanism we call 'facilitation in the halo'. Most coexisting species are displaced to an interphase area between patches, where two opposite forces reach equilibrium: attenuated severe conditions by proximity to the F. curvifolia canopy (nutrient-rich islands) and competitive exclusion mitigated by avoiding direct contact with F. curvifolia.

Age regression from soft aligned face images using low computational resources

The initial step in most facial age estimation systems consists of accurately aligning a model to the output of a face detector (e.g. an Active Appearance Model). This fitting process is very expensive in terms of computational resources and prone to get stuck in local minima. This makes it impractical for analysing faces in resource limited computing devices. In this paper we build a face age regressor that is able to work directly on faces cropped using a state-of-the-art face detector. Our procedure uses K nearest neighbours (K-NN) regression with a metric based on a properly tuned Fisher Linear Discriminant Analysis (LDA) projection matrix. On FG-NET we achieve a state-of-the-art Mean Absolute Error (MAE) of 5.72 years with manually aligned faces. Using face images cropped by a face detector we get a MAE of 6.87 years in the same database. Moreover, most of the algorithms presented in the literature have been evaluated on single database experiments and therefore, they report optimistically biased results. In our cross-database experiments we get a MAE of roughly 12 years, which would be the expected performance in a real world application.

Nup93, a Vertebrate Homologue of Yeast Nic96p, Forms a Complex with a Novel 205-kDa Protein and Is Required for Correct Nuclear Pore Assembly

Yeast and vertebrate nuclear pores display significant morphological similarity by electron microscopy, but sequence similarity between the respective proteins has been more difficult to observe. Herein we have identified a vertebrate nucleoporin, Nup93, in both human and Xenopus that has proved to be an evolutionarily related homologue of the yeast nucleoporin Nic96p. Polyclonal antiserum to human Nup93 detects corresponding proteins in human, rat, and Xenopus cells. Immunofluorescence and immunoelectron microscopy localize vertebrate Nup93 at the nuclear basket and at or near the nuclear entry to the gated channel of the pore. Immunoprecipitation from both mammalian and Xenopus cell extracts indicates that a small fraction of Nup93 physically interacts with the nucleoporin p62, just as yeast Nic96p interacts with the yeast p62 homologue. However, a large fraction of vertebrate Nup93 is extracted from pores and is also present in Xenopus egg extracts in complex with a newly discovered 205-kDa protein. Mass spectrometric sequencing of the human 205-kDa protein reveals that this protein is encoded by an open reading frame, KIAAO225, present in the human database. The putative human nucleoporin of 205 kDa has related sequence homologues in Caenorhabditis elegans and Saccharomyces cerevisiae. To analyze the role of the Nup93 complex in the pore, nuclei were assembled that lack the Nup93 complex after immunodepletion of a Xenopus nuclear reconstitution extract. The Nup93-complex–depleted nuclei are clearly defective for correct nuclear pore assembly. From these experiments, we conclude that the vertebrate and yeast pore have significant homology in their functionally important cores and that, with the identification of Nup93 and the 205-kDa protein, we have extended the knowledge of the nearest-neighbor interactions of this core in both yeast and vertebrates.

Non-CpG methylation is prevalent in embryonic stem cells and may be mediated by DNA methyltransferase 3a

Current evidence indicates that methylation of cytosine in mammalian DNA is restricted to both strands of the symmetrical sequence CpG, although there have been sporadic reports that sequences other than CpG may also be methylated. We have used a dual-labeling nearest neighbor technique and bisulphite genomic sequencing methods to investigate the nearest neighbors of 5-methylcytosine residues in mammalian DNA. We find that embryonic stem cells, but not somatic tissues, have significant cytosine-5 methylation at CpA and, to a lesser extent, at CpT. As the expression of the de novo methyltransferase Dnmt3a correlates well with the presence of non-CpG methylation, we asked whether Dnmt3a might be responsible for this modification. Analysis of genomic methylation in transgenic Drosophila expressing Dnmt3a reveals that Dnmt3a is predominantly a CpG methylase but also is able to induce methylation at CpA and at CpT.

Prediction of the stability of DNA triplexes.

We present rules that allow one to predict the stability of DNA pyrimidine.purine.pyrimidine (Y.R.Y) triple helices on the basis of the sequence. The rules were derived from van't Hoff analysis of 23 oligonucleotide triplexes tested at a variety of pH values. To predict the enthalpy of triplex formation (delta H degrees), a simple nearest-neighbor model was found to be sufficient. However, to accurately predict the free energy of the triplex (delta G degrees), a combination model consisting of five parameters was needed. These parameters were (i) the delta G degrees for helix initiation, (ii) the delta G degrees for adding a T-A.T triple, (iii) the delta G degrees for adding a C(+)-G.C triple, (iv) the penalty for adjacent C bases, and (v) the pH dependence of the C(+)-G.C triple's stability. The fitted parameters are highly consistent with thermodynamic data from the basis set, generally predicting both delta H degrees and delta G degrees to within the experimental error. Examination of the parameters points out several interesting features. The combination model predicts that C(+) -G.C. triples are much more stabilizing than T-A.T triples below pH 7.0 and that the stability of the former increases approximately equal to 1 kcal/mol per pH unit as the pH is decreased. Surprisingly though, the most stable sequence is predicted to be a CT repeat, as adjacent C bases partially cancel the stability of one another. The parameters successfully predict tm values from other laboratories, with some interesting exceptions.

How are close residues of protein structures distributed in primary sequence?

Structurally neighboring residues are categorized according to their separation in the primary sequence as proximal (1-4 positions apart) and otherwise distal, which in turn is divided into near (5-20 positions), far (21-50 positions), very far ( > 50 positions), and interchain (from different chains of the same structure). These categories describe the linear distance histogram (LDH) for three-dimensional neighboring residue types. Among the main results are the following: (i) nearest-neighbor hydrophobic residues tend to be increasingly distally separated in the linear sequence, thus most often connecting distinct secondary structure units. (ii) The LDHs of oppositely charged nearest-neighbors emphasize proximal positions with a subsidiary maximum for very far positions. (iii) Cysteine-cysteine structural interactions rarely involve proximal positions. (iv) The greatest numbers of interchain specific nearest-neighbors in protein structures are composed of oppositely charged residues. (v) The largest fraction of side-chain neighboring residues from beta-strands involves near positions, emphasizing associations between consecutive strands. (vi) Exposed residue pairs are predominantly located in proximal linear positions, while buried residue pairs principally correspond to far or very far distal positions. The results are principally invariant to protein sizes, amino acid usages, linear distance normalizations, and over- and underrepresentations among nearest-neighbor types. Interpretations and hypotheses concerning the LDHs, particularly those of hydrophobic and charged pairings, are discussed with respect to protein stability and functionality. The pronounced occurrence of oppositely charged interchain contacts is consistent with many observations on protein complexes where multichain stabilization is facilitated by electrostatic interactions.

Temporal fluctuations in coherence of brain waves.

As a measure of dynamical structure, short-term fluctuations of coherence between 0.3 and 100 Hz in the electroencephalogram (EEG) of humans were studied from recordings made by chronic subdural macroelectrodes 5-10 mm apart, on temporal, frontal, and parietal lobes, and from intracranial probes deep in the temporal lobe, including the hippocampus, during sleep, alert, and seizure states. The time series of coherence between adjacent sites calculated every second or less often varies widely in stability over time; sometimes it is stable for half a minute or more. Within 2-min samples, coherence commonly fluctuates by a factor up to 2-3, in all bands, within the time scale of seconds to tens of seconds. The power spectrum of the time series of these fluctuations is broad, extending to 0.02 Hz or slower, and is weighted toward the slower frequencies; little power is faster than 0.5 Hz. Some records show conspicuous swings with a preferred duration of 5-15s, either irregularly or quasirhythmically with a broad peak around 0.1 Hz. Periodicity is not statistically significant in most records. In our sampling, we have not found a consistent difference between lobes of the brain, subdural and depth electrodes, or sleeping and waking states. Seizures generally raise the mean coherence in all frequencies and may reduce the fluctuations by a ceiling effect. The coherence time series of different bands is positively correlated (0.45 overall); significant nonindependence extends for at least two octaves. Coherence fluctuations are quite local; the time series of adjacent electrodes is correlated with that of the nearest neighbor pairs (10 mm) to a coefficient averaging approximately 0.4, falling to approximately 0.2 for neighbors-but-one (20 mm) and to < 0.1 for neighbors-but-two (30 mm). The evidence indicates fine structure in time and space, a dynamic and local determination of this measure of cooperativity. Widely separated frequencies tending to fluctuate together exclude independent oscillators as the general or usual basis of the EEG, although a few rhythms are well known under special conditions. Broad-band events may be the more usual generators. Loci only a few millimeters apart can fluctuate widely in seconds, either in parallel or independently. Scalp EEG coherence cannot be predicted from subdural or deep recordings, or vice versa, and intracortical microelectrodes show still greater coherence fluctuation in space and time. Widely used computations of chaos and dimensionality made upon data from scalp or even subdural or depth electrodes, even when reproducible in successive samples, cannot be considered representative of the brain or the given structure or brain state but only of the scale or view (receptive field) of the electrodes used. Relevant to the evolution of more complex brains, which is an outstanding fact of animal evolution, we believe that measures of cooperativity are likely to be among the dynamic features by which major evolutionary grades of brains differ.

Electrical investigation of p-type 4H-SiC heavily doped by ion implantation

In the last decades, an increasing interest in the research field of wide bandgap semiconductors was observed, mostly due to the progressive approaching of silicon-based devices to their theoretical limits. 4H-SiC is an example among these, and is a mature compound for applications. The main advantages offered 4H-SiC in comparison with silicon are an higher breakdown field, an higher thermal conductivity, a higher operating temperature, very high hardness and melting point, biocompatibility, but also low switching losses in high frequencies applications and lower on-resistances in unipolar devices. Then, 4H-SiC power devices offer great performance improvement; moreover, they can work in hostile environments where silicon power devices cannot function. Ion implantation technology is a key process in the fabrication of almost all kinds of SiC devices, owing to the advantage of a spatially selective doping. This work is dedicated to the electrical investigation of several differently-processed 4H-SiC ion- implanted samples, mainly through Hall effect and space charge spectroscopy experiments. It was also developed the automatic control (Labview) of several experiments. In the work, the effectiveness of high temperature post-implant thermal treatments (up to 2000°C) were studied and compared considering: (i) different methods, (ii) different temperatures and (iii) different duration of the annealing process. Preliminary p + /n and Schottky junctions were also investigated as simple test devices. 1) Heavy doping by ion implantation of single off-axis 4H-SiC layers The electrical investigation is one of the most important characterization of ion-implanted samples, which must be submitted to mandatory post-implant thermal treatment in order to both (i) recover the lattice after ion bombardment, and (ii) address the implanted impurities into lattice sites so that they can effectively act as dopants. Electrical investigation can give fundamental information on the efficiency of the electrical impurity activation. To understand the results of the research it should be noted that: (a) To realize good ohmic contacts it is necessary to obtain spatially defined highly doped regions, which must have conductivity as low as possible. (b) It has been shown that the electrical activation efficiency and the electrical conductivity increase with the annealing temperature increasing. (c) To maximize the layer conductivity, temperatures around 1700°C are generally used and implantation density high till to 10 21 cm -3 . In this work, an original approach, different from (c), is explored by the using very high annealing temperature, around 2000°C, on samples of Al + -implant concentration of the order of 10 20 cm -3 . Several Al + -implanted 4H-SiC samples, resulting of p-type conductivity, were investigated, with a nominal density varying in the range of about 1-5∙10 20 cm -3 and subjected to two different high temperature thermal treatments. One annealing method uses a radiofrequency heated furnace till to 1950°C (Conventional Annealing, CA), the other exploits a microwave field, providing a fast heating rate up to 2000°C (Micro-Wave Annealing, MWA). In this contest, mainly ion implanted p-type samples were investigated, both off-axis and on-axis <0001> semi-insulating 4H-SiC. Concerning p-type off-axis samples, a high electrical activation of implanted Al (50-70%) and a compensation ratio below 10% were estimated. In the work, the main sample processing parameters have been varied, as the implant temperature, CA annealing duration, and heating/cooling rates, and the best values assessed. MWA method leads to higher hole density and lower mobility than CA in equivalent ion implanted layers, resulting in lower resistivity, probably related to the 50°C higher annealing temperature. An optimal duration of the CA treatment was estimated in about 12-13 minutes. A RT resistivity on the lowest reported in literature for this kind of samples, has been obtained. 2) Low resistivity data: variable range hopping Notwithstanding the heavy p-type doping levels, the carrier density remained less than the critical one required for a semiconductor to metal transition. However, the high carrier densities obtained was enough to trigger a low temperature impurity band (IB) conduction. In the heaviest doped samples, such a conduction mechanism persists till to RT, without significantly prejudice the mobility values. This feature can have an interesting technological fall, because it guarantee a nearly temperature- independent carrier density, it being not affected by freeze-out effects. The usual transport mechanism occurring in the IB conduction is the nearest neighbor hopping: such a regime is effectively consistent with the resistivity temperature behavior of the lowest doped samples. In the heavier doped samples, however, a trend of the resistivity data compatible with a variable range hopping (VRH) conduction has been pointed out, here highlighted for the first time in p-type 4H-SiC. Even more: in the heaviest doped samples, and in particular, in those annealed by MWA, the temperature dependence of the resistivity data is consistent with a reduced dimensionality (2D) of the VRH conduction. In these samples, TEM investigation pointed out faulted dislocation loops in the basal plane, whose average spacing along the c-axis is comparable with the optimal length of the hops in the VRH transport. This result suggested the assignment of such a peculiar behavior to a kind of spatial confinement into a plane of the carrier hops. 3) Test device the p + -n junction In the last part of the work, the electrical properties of 4H-SiC diodes were also studied. In this case, a heavy Al + ion implantation was realized on n-type epilayers, according to the technological process applied for final devices. Good rectification properties was shown from these preliminary devices in their current-voltage characteristics. Admittance spectroscopy and deep level transient spectroscopy measurements showed the presence of electrically active defects other than the dopants ones, induced in the active region of the diodes by ion implantation. A critical comparison with the literature of these defects was performed. Preliminary to such an investigation, it was assessed the experimental set up for the admittance spectroscopy and current-voltage investigation and the automatic control of these measurements.

Janus: an FPGA-based system for high-performance scientific computing

This paper describes JANUS, a modular massively parallel and reconfigurable FPGA-based computing system. Each JANUS module has a computational core and a host. The computational core is a 4x4 array of FPGA-based processing elements with nearest-neighbor data links. Processors are also directly connected to an I/O node attached to the JANUS host, a conventional PC. JANUS is tailored for, but not limited to, the requirements of a class of hard scientific applications characterized by regular code structure, unconventional data manipulation instructions and not too large data-base size. We discuss the architecture of this configurable machine, and focus on its use on Monte Carlo simulations of statistical mechanics. On this class of application JANUS achieves impressive performances: in some cases one JANUS processing element outperfoms high-end PCs by a factor ≈1000. We also discuss the role of JANUS on other classes of scientific applications.

Quantum simulation of a topological Mott insulator with Rydberg atoms in a Lieb lattice

We propose a realistic scheme to quantum simulate the so-far experimentally unobserved topological Mott insulator phase-an interaction-driven topological insulator-using cold atoms in an optical Lieb lattice. To this end, we study a system of spinless fermions in a Lieb lattice, exhibiting repulsive nearest-and next-to-nearest-neighbor interactions and derive the associated zero-temperature phase diagram within mean-field approximation. In particular, we analyze how the interactions can dynamically generate a charge density wave ordered, a nematic, and a topologically nontrivial quantum anomalous Hall phase. We characterize the topology of the different phases by the Chern number and discuss the possibility of phase coexistence. Based on the identified phases, we propose a realistic implementation of this model using cold Rydberg-dressed atoms in an optical lattice. The scheme, which allows one to access, in particular, the topological Mott insulator phase, robustly and independently of its exact position in parameter space, merely requires global, always-on off-resonant laser coupling to Rydberg states and is feasible with state-of-the-art experimental techniques that have already been demonstrated in the laboratory.

On the suitability of Prototype Selection methods for kNN classification with distributed data

In the current Information Age, data production and processing demands are ever increasing. This has motivated the appearance of large-scale distributed information. This phenomenon also applies to Pattern Recognition so that classic and common algorithms, such as the k-Nearest Neighbour, are unable to be used. To improve the efficiency of this classifier, Prototype Selection (PS) strategies can be used. Nevertheless, current PS algorithms were not designed to deal with distributed data, and their performance is therefore unknown under these conditions. This work is devoted to carrying out an experimental study on a simulated framework in which PS strategies can be compared under classical conditions as well as those expected in distributed scenarios. Our results report a general behaviour that is degraded as conditions approach to more realistic scenarios. However, our experiments also show that some methods are able to achieve a fairly similar performance to that of the non-distributed scenario. Thus, although there is a clear need for developing specific PS methodologies and algorithms for tackling these situations, those that reported a higher robustness against such conditions may be good candidates from which to start.