1,965 research outputs found
Diseño e implementación en hardware reconfigurable de un sistema de reconocimiento de gestos de la mano basado en visión por computador
No full text[SPA] En esta tesis se proponen un sistema para el reconocimiento de gestos de la mano basado en visión por computador y el diseño de su implementación hardware. El propósito del reconocimiento de gestos es proporcionar a una computadora la capacidad de detectar gestos realizados por una persona. Esta tarea, innata para el ser humano, ha resultado ser compleja y difícil de automatizar. Entre las diferentes aproximaciones al problema, una de las principales líneas de trabajo es el empleo de visión por computador. El desarrollo de las técnicas de procesamiento en visión por computador ha proporcionado herramientas para que sistemas basados en microprocesadores analicen imágenes adquiridas por cámaras e intenten extraer de ellas información que resulte de interés para cualquier aplicación. Analizado desde esta perspectiva, el reconocimiento de los gestos de la mano es un problema de reconocimiento de objetos, campo en el que se distinguen dos niveles: de instancia, cuando se busca un objeto específico, una persona concreta; y de categoría, cuando se pretende reconocer cualquier instancia de un tipo de objeto. Este segundo nivel persigue, definida una colección de categorías de objetos y dada una imagen, determinar si hay algún objeto de una categoría presente en ella. En particular, en esta tesis la categoría es un gesto, definido por una determinada posición y orientación de la mano y por la configuración de los dedos. En este marco, se ha definido una colección de categorías —una biblioteca de gestos— que se pretende reconocer y, con tal objetivo, se ha desarrollado un conjunto de etapas de procesamiento y algoritmos que conforman el sistema de reconocimiento de gestos de la mano. En primer lugar, se pretende separar la mano del resto de la imagen. Para ello, se propone un algoritmo de reconocimiento del color de la piel, basado en modelos construidos en diferentes espacios de color. Desarrollado con el propósito mencionado, puede resultar de interés en cualquiera de las numerosas aplicaciones en las que se lleva a cabo segmentación de imágenes basada en el color de la piel. Una vez segmentada la imagen, se propone detectar la mano y reconocer el gesto identificando sus partes elementales —palma y dedos— por medio de la convolución bidimensional de la imagen segmentada con un conjunto de plantillas definidas con tal fin. A partir del análisis de la información resultante de las convoluciones de estas plantillas con las imágenes de una base de datos de gestos creada con este propósito, se ha construido un modelo para cada uno de los gestos de la biblioteca. En el proceso de desarrollo de las diferentes etapas, la metodología de diseño ha buscado favorecer la modularidad y la escalabilidad suficiente como para posibilitar la actualización de la biblioteca de gestos y la adaptación del funcionamiento global del sistema a diversas aplicaciones. Para proporcionar al usuario una experiencia satisfactoria en el manejo del sistema de reconocimiento es imprescindible que la interacción se realice con la mayor naturalidad. Esto requiere que el usuario perciba que el sistema responde de manera inmediata a sus acciones, lo cual implica que la rapidez de respuesta del sistema sea una prestación clave. Con el propósito de optimizar las prestaciones temporales de la ejecución de los algoritmos de procesamiento, se han propuesto soluciones para su implementación en hardware reconfigurable. Los dispositivos FPGA son una plataforma muy adecuada para acelerar algoritmos de alta carga computacional. Su estructura interna los hace ideales para explotar el paralelismo a nivel de píxel inherente a los algoritmos de procesamiento de imagen de bajo nivel, también el paralelismo a nivel de instrucción por medio de la segmentación de cauce y, al mismo tiempo, el paralelismo a más alto nivel para ejecutar simultáneamente distintas operaciones. Por todo ello, las FPGA son la plataforma hardware adecuada para la implementación de nuestro sistema. Empleando dispositivos y herramientas de Xilinx R, se ha diseñado, implementado y validado un sistema digital que ejecuta las tareas de procesamiento involucradas en el reconocimiento de los gestos, en el marco de una arquitectura híbrida hardware/software. El criterio de particionado ha sido la escala temporal de las tareas, en la que se distinguen dos niveles: nivel de píxel y nivel de imagen. Para resoluciones y sensores de imagen típicos de sistemas embebidos, los algoritmos que operan con los valores de los píxeles lo hacen en el orden de los nanosegundos. Su dominio propio es el del hardware, donde es posible explotar el paralelismo de las operaciones y la flexibilidad de la arquitectura de las FPGA para lograr procesamiento en tiempo real. Por su parte, las tareas a nivel de imagen, en el orden de los milisegundos, conviene que se ejecuten en software. Dentro del sistema digital diseñado, en esta tesis se desarrollan soluciones para la implementación hardware de las dos tareas a nivel de píxel más relevantes: la segmentación según el color de la piel y la convolución bidimensional. En particular, para la convolución, que es la etapa con mayor carga computacional, se proponen arquitecturas tanto para la realización de las operaciones implicadas en su cálculo como para el almacenamiento temporal de los datos. Los resultados obtenidos en las diferentes campañas de test demuestran tanto la bondad de la solución propuesta al problema planteado como la viabilidad de su implementación por medio de los dispositivos FPGA. [ENG] In this thesis, a system for hand gesture recognition based on computer vision and the design of its hardware implementation are proposed. The purpose of gesture recognition is to provide a computer with the ability to detect gestures made by a person. This task, innate to humans, has proven to be complex and difficult to automate. Among the different approaches to the problem, one of the main lines of work is the use of computer vision. The development of computer vision processing techniques has provided tools for microprocessor–based systems to analyze images acquired by cameras and try to extract from them information of interest for any application. Analyzed from this perspective, hand gesture recognition is an object recognition problem, a field in which two levels can be distinguished: instance level, when looking for a specific object, a specific person; and category level, when trying to recognize any instance of a type of object. This second level aims, when a collection of object categories is defined and given an image, to determine if there is any object of a category present in it. In particular, in this thesis the category is a gesture, defined by a certain position and orientation of the hand and by the configuration of the fingers. In this framework, a collection of categories —a gesture library— that is intended to be recognized has been defined and, with such an objective, a set of processing steps and algorithms that make up the hand gesture recognition system has been developed. First, it is intended to separate the hand from the rest of the image. For this purpose, a skin color recognition algorithm is proposed, based on models built in different color spaces. Developed for the aforementioned purpose, it may be of interest in any of the numerous applications where skin color–based image segmentation is carried out. Once the image is segmented, it is proposed to detect the hand and recognize the gesture by identifying its elementary parts—palm and fingers—by means of two–dimensional convolution of the segmented image with a set of templates defined for that purpose. From the analysis of the information resulting from the convolutions of these templates with the images of a gesture database, a model has been constructed for each of the gestures in the library. In the development process of the different stages, the design methodology has sought to favor modularity and scalability sufficient to enable the updating of the gesture library and the adaptation of the overall functioning of the system to different applications. In order to provide the user with a satisfactory experience in the operation of the recognition system, it is essential that the interaction is carried out as naturally as possible. This requires that the user perceives that the system responds immediately to his or her actions, which implies that the speed of the system’s response is a key performance indicator. In order to optimize the temporal performance of the execution of processing algorithms, solutions based on reconfigurable hardware were explored. FPGA devices are a suitable platform for accelerating computationally intensive algorithms. Their internal structure makes them ideal for exploiting the pixel–level parallelism inherent in low–level image processing algorithms, also instruction–level parallelism through pipeline segmentation and, at the same time, higher–level parallelism for the simultaneous execution of different operations. For all these reasons, FPGAs are the proper hardware platform for the implementation of our system. Using Xilinx R devices and tools, we have designed, implemented, and validated a digital system that executes the processing tasks involved in gesture recognition, in the framework of a hybrid hardware/software architecture. The partitioning criterion has been the time scale of the tasks, in which two levels are distinguished: pixel level and image level. For resolutions and image sensors typical of embedded systems, the algorithms that operate on pixel values do so on the order of nanoseconds. Their home domain is hardware, where it is possible to exploit the parallelism of operations and the flexibility of the FPGA architecture to achieve real–time processing. On the other hand, image–level tasks, in the order of milliseconds, should be executed in software. Within the designed digital system, this thesis develops solutions for the hardware implementation of the two most relevant pixel–level tasks: skin color segmentation and two– dimensional convolution. In particular, for convolution, which is the most computationally intensive step, architectures are proposed both for the performance of the operations involved in its computation and for the temporal storage of the data. The results obtained in the different test campaigns demonstrate both the goodness of the proposed solution to the computer vision problem and the feasibility of its implementation by means of FPGA devices.Escuela Internacional de Doctorado de la Universidad Politécnica de CartagenaPrograma Doctorado en Tecnologías IndustrialesUniversidad Politécnica de Cartagen
Filamentarity and inhomogeneities in low dimensional superconductors
No full textIn this Thesis, I investigate the problem of the interplay between disorder and low dimensionality in superconductors.
From the microscopic point of view, I show that the presence of impurities in the superconducting condensate can produce a pairbreaking effect at the Lifshitz transition in multibands superconductors, avoiding or at least circumventing Anderson’s Theorem. This is consistent with the observed suppression of the superconducting critical temperature Tc in SrTiO3 -based heterostructures as a function of the gate potential Vg . This study allows us to disentangle microscopic from mesoscopic disorder in SrTiO3 -based interfaces: microscopic impurities are in fact necessary to explain the suppression of Tc observed when multiband superconductivity is involved; the global behavior is instead well captured only if the strongly inhomogeneous nature of such compounds is considered.
Indeed, disorder can also appear on a mesoscopic length scale. The electronic condensate can in fact segregate into regions large enough to define a local phase but small compared with the sample. The reasons behind this inhomogeneity of the superconducting order parameter can be several, depending on the system in exam: in oxides heterostructures it may be connected to the thermodynamic instability caused by the electrostatic potential confining the electron gas at the interface, in transition metal dichalcogenides the instability may be introduced by the gating ionic liquid, whereas sometimes the phase separation can be caused by the presence of microscopic impurities in addition with the competition of superconductivity with another order parameter, e.g., the charge ordering in the case of cuprates. Such a phase separation may appear as a filamentary structure.
Concerning mesoscopic inhomogeneities, I will focus on two different materials.
In SrTiO3 -based eterostructures, where filamentary superconductivity is embedded in a metallic matrix, I study the consequences of filamentarity in transport properties, assuming a priori a fractal-like organization of the electronic condensate and calculating the complex conductance with a Random Impedance Network model.
The geometry of the filamentary structure plays a crucial role, especially in the behavior of the superfluid stiffness as a function of the temperature. Although the motivation of this work is connected to resonant microwave experiments performed on a LaAlO3 /SrTiO3 interface, the results are quite general and can in principle be
applied to other materials displaying filamentary superconductivity.
In cuprates, I focus on the phase competition as the most likely reason of filamentarity. The charge ordering-superconducting (CO-SC) competition is studied by means of Monte Carlo simulations within an anisotropic Heisenberg model accounting for the basic physical symmetries involved, the out-of-plane pseudospin component
mapping two possible charge density waves (CDW) variants while the in-plane component standing for the SC order parameter. The anisotropy term α is taken as the control parameter, tuning the transition from Berezinskii-Kosterlitz-Thouless (BKT) to the charge ordered state. The phase diagram T c vs α is studied both in the clean case and in a random field, the presence of microscopic impurities being necessary to stabilize the clustering of charge
ordered domains and the appearance of filamentary superconductivity
Forging the Forger: An Attempt to Improve Authorship Verification via Data Augmentation
No full textAuthorship Verification (AV) is a text classification task concerned with inferring whether a candidate text has been written by one specific author (A) or by someone else ( A ̄ ̄ ̄ ̄ ). Itehas been shown that many AV systems are vulnerable to adversarial attacks, where a malicious author actively tries to fool the classifier by either concealing their writing style, oreby imitating the style of another author. Inethis paper, weeinvestigate the potential benefits of augmenting the classifier training set with (negative) synthetic examples. These synthetic examples are generated to imitate the style of A. Weeanalyze the improvements in the classifier predictions that this augmentation brings to bear in the task of AV in an adversarial setting. Ineparticular, weeexperiment with three different generator architectures (one based on Recurrent Neural Networks, another based on small-scale transformers, and another based on the popular GPT model) and with two training strategies (one inspired by standard Language Models, and another inspired by Wasserstein Generative Adversarial Networks). Weeevaluate our hypothesis on five datasets (three of which have been specifically collected to represent an adversarial setting) and using two learning algorithms for the AV classifier (Support Vector Machines and Convolutional Neural Networks). This experimentation yields negative results, revealing that, although our methodology proves effective in many adversarial settings, its benefits are too sporadic for a pragmatical application
Measuring Fairness Under Unawareness of Sensitive Attributes: A Quantification-Based Approach
No full textAlgorithms and models are increasingly deployed to inform decisions about
people, inevitably affecting their lives. As a consequence, those in charge of
developing these models must carefully evaluate their impact on different
groups of people and favour group fairness, that is, ensure that groups
determined by sensitive demographic attributes, such as race or sex, are not
treated unjustly. To achieve this goal, the availability (awareness) of these
demographic attributes to those evaluating the impact of these models is
fundamental. Unfortunately, collecting and storing these attributes is often in
conflict with industry practices and legislation on data minimisation and
privacy. For this reason, it can be hard to measure the group fairness of
trained models, even from within the companies developing them. In this work,
we tackle the problem of measuring group fairness under unawareness of
sensitive attributes, by using techniques from quantification, a supervised
learning task concerned with directly providing group-level prevalence
estimates (rather than individual-level class labels). We show that
quantification approaches are particularly suited to tackle the
fairness-under-unawareness problem, as they are robust to inevitable
distribution shifts while at the same time decoupling the (desirable) objective
of measuring group fairness from the (undesirable) side effect of allowing the
inference of sensitive attributes of individuals. More in detail, we show that
fairness under unawareness can be cast as a quantification problem and solved
with proven methods from the quantification literature. We show that these
methods outperform previous approaches to measure demographic parity in five
experimental protocols, corresponding to important challenges that complicate
the estimation of classifier fairness under unawareness.Comment: Accepted for publication in the Journal of Artificial Intelligence
Researc
Learning to Quantify
No full textThis open access book provides an introduction and an overview of learning to quantify (a.k.a. “quantification”), i.e. the task of training estimators of class proportions in unlabeled data by means of supervised learning. In data science, learning to quantify is a task of its own related to classification yet different from it, since estimating class proportions by simply classifying all data and counting the labels assigned by the classifier is known to often return inaccurate (“biased”) class proportion estimates. The book introduces learning to quantify by looking at the supervised learning methods that can be used to perform it, at the evaluation measures and evaluation protocols that should be used for evaluating the quality of the returned predictions, at the numerous fields of human activity in which the use of quantification techniques may provide improved results with respect to the naive use of classification techniques, and at advanced topics in quantification research. The book is suitable to researchers, data scientists, or PhD students, who want to come up to speed with the state of the art in learning to quantify, but also to researchers wishing to apply data science technologies to fields of human activity (e.g., the social sciences, political science, epidemiology, market research) which focus on aggregate (“macro”) data rather than on individual (“micro”) data
The use of ozone therapy for treatment of periodontal disease: a split-mouth, randomized, controlled clinical trial
No full textPeriodontal treatment has the aim to reduce oral infection and prevent the progression of the disease. The potential benefits of new therapy with Ozonline® for periodontal treatment, include improved patient compliance and an easier access to periodontal pocket. The objective of this study was to explore the efficacy of Ozonline® in the treatment of chronic periodontitis in adult patients. A randomized controlled split-mouth study was carried out in ten patients (5 men and 5 women age 42-73 mean 55 ±7) with a diagnosis of chronic periodontitis. None of these patients received any surgical or non-surgical periodontal therapy and demonstrated radiographic evidence of moderate bone loss. The mouth has been divided into upper right and left quadrants. The upper and lower right quadrants were treated with ultrasonic scaler, the left quadrants with ultrasonic scaler with ozonated water (Ozonline®). 10 microbiological samples were collected from upper left quadrants and 10 from upper right quadrants from each patient. Microbiological samples were collected from the sites of the patients at baseline and at the 7th day. 20 localized chronic periodontitis sites were selected (10 in left quadrants and 10 in right quadrants). After the treatment with Ozonline®, a remarkable decrease in bacteria amount, both for some species and for the total count was observed in the left quadrants respect to right ones. Specifically, T. forsythia and T. denticola were eradicated whereas Total Bacteria Loading and Fusobacterium nucleatum showed a reduction of 38% and 55%, respect to right quadrants. Our study demonstrated the efficacy of the Ozonline® in the management of moderate to severe chronic periodontitis.
The use of ozone therapy for treatment of periodontal disease: A split-mouth, randomized, controlled clinical trial
No full textPeriodontal treatment has the aim to reduce oral infection and prevent the progression of the disease. The potential benefits of new therapy with Ozonline® for periodontal treatment, include improved patient compliance and an easier access to periodontal pocket. The objective of this study was to explore the efficacy of Ozonline® in the treatment of chronic periodontitis in adult patients. A randomized controlled split-mouth study was carried out in ten patients (5 men and 5 women age 42-73 mean 55 ± 7) with a diagnosis of chronic periodontitis. None of these patients received any surgical or non-surgical periodontal therapy and demonstrated radiographic evidence of moderate bone loss. The mouth has been divided into upper right and left quadrants. The upper and lower right quadrants were treated with ultrasonic scaler, the left quadrants with ultrasonic scaler with ozonated water (Ozonline®). 10 microbiological samples were collected from upper left quadrants and 10 from upper right quadrants from each patient. Microbiological samples were collected from the sites of the patients at baseline and at the 7th day. 20 localized chronic periodontitis sites were selected (10 in left quadrants and 10 in rigt quadrants). After the treatment with Ozonline®, a remarkable decrease in bacteria amount, both for some species and for the total count was observed in the left quadrants respect to right ones. Specifically, T. forsythia and T. denticola were eradicated whereas Total Bacteria Loading and Fusobacterium Nucleatum showed a reduction of 38% and 55%, respect to right quadrants. Our study demonstrated the efficacy of the Ozonline® in the management of moderate to severe chronic periodontitis
: Two Datasets for the Computational Authorship Analysis of Medieval Latin Texts
No full textWe present and make available MedLatinEpi and MedLatinLit, two datasets of medieval Latin texts to be used in research
on computational authorship analysis. MedLatinEpi and MedLatinLit consist of 294 and 30 curated texts, respectively,
labelled by author; MedLatinEpi texts are of epistolary nature, while MedLatinLit texts consist of literary comments and
treatises about various subjects. As such, these two datasets lend themselves to supporting research in authorship analysis
tasks, such as authorship attribution, authorship verification, or same-author verification. Along with the datasets, we provide
experimental results, obtained on these datasets, for the authorship verification task, i.e., the task of predicting whether a
text of unknown authorship was written by a candidate author. We also make available the source code of the authorship
verification system we have used, thus allowing our experiments to be reproduced, and to be used as baselines, by other
researchers. We also describe the application of the above authorship verification system, using these datasets as training
data, for investigating the authorship of two medieval epistles whose authorship has been disputed by scholars.
on computational authorship analysis. MedLatinEpi and MedLatinLit consist of 294 and 30 curated texts, respectively,
labelled by author; MedLatinEpi texts are of epistolary nature, while MedLatinLit texts consist of literary comments and
treatises about various subjects. As such, these two datasets lend themselves to supporting research in authorship analysis
tasks, such as authorship attribution, authorship verification, or same-author verification. Along with the datasets, we provide
experimental results, obtained on these datasets, for the authorship verification task, i.e., the task of predicting whether a
text of unknown authorship was written by a candidate author. We also make available the source code of the authorship
verification system we have used, thus allowing our experiments to be reproduced, and to be used as baselines, by other
researchers. We also describe the application of the above authorship verification system, using these datasets as training
data, for investigating the authorship of two medieval epistles whose authorship has been disputed by scholars
Syllabic quantity patterns as rhythmic features for Latin authorship attribution
No full textIt is well known that, within the Latin production of written text, peculiar metric schemes were followed not only in poetic compositions, but also in many prose works. Such metric patterns were based on so-called syllabic quantity, that is, on the length of the involved syllables, and there is substantial evidence suggesting that certain authors had a preference for certain metric patterns over others. In this research we investigate the possibility to employ syllabic quantity as a base for deriving rhythmic features for the task of computational authorship attribution of Latin prose texts. We test the impact of these features on the authorship attribution task when combined with other topic-agnostic features. Our experiments, carried out on three different datasets using support vector machines (SVMs) show that rhythmic features based on syllabic quantity are beneficial in discriminating among Latin prose authors
Drug-induced gingival overgrowth: The effect of cyclosporin a and mycophenolate mophetil on human gingival fibroblasts
No full textDrug-induced gingival overgrowth may occur after a chronic administration of three classes of systemic drugs: Anticonvulsants, immunosuppressants, and calcium channel blockers. This study aimed to investigate how cyclosporin A and mycophenolate mophetil (immunosuppressive drugs) could interfere with human gingival fibroblasts functions, leading to gingival enlargement. Human gingival fibroblasts derived fromthe tissue of a 60-year-old female were cultured in aDMEMEmedium. A stock solution with 1 mg/mL of mycophenolate and 1 mg/mL of cyclosporine were prepared and dissolved in a DMEM medium to prepare a serial dilution at the concentrations of 5000, 2000, 1000, 500, and 100 ng/mL, for both treatments. Cell viability was measured using the PrestoBlueTM Reagent Protocol. Quantitative real-time RT-PCR was performed in order to analyze the expression of 57 genes coding for gingival fibroblasts "Extracellular Matrix and Adhesion Molecules". Mycophenolate and cyclosporine had no effect on fibroblast cell viability at 1000 ng/mL. Both the treatments showed similar effects on the expression profiling of treated cells: Downregulation of most extracellular matrix metalloproteases genes (MMP8, MMP11, MMP15, MMP16, MMP24) was assessed, while CDH1, ITGA2, ITGA7, LAMB3, MMP12, and MMP13 were recorded to be upregulated in fibroblasts treated with immunosuppressive drugs. It has been demonstrated that gingival overgrowth can be caused by the chronic administration of cyclosporin A and mycophenolate mophetil. However, given the contrasting data of literature, further investigations are needed, making clear the possible effects of immunosuppressive drugs on fibroblasts
- …
