Desarrollo de interfaces de usuario para una herramienta de telecontrol y automatización industrial

[ES] Este proyecto de fin de carrera aborda la actualización y refactorización de la aplicación Hecaton. Esta aplicación permite la monitorización y actuación en instalaciones industriales de manera remota a través de un interfaz web. Para ello hace uso de sensores y actuadores que, conectados a través de un equipo de adquisición de datos a un sistema informático servidor, permiten obtener, manipular y almacenar los datos y eventos recibidos. Hecaton ha sido desarrollado enteramente utilizando software libre. Además, el sistema permite ser personalizado, lo que posibilita su uso en todo tipo de escenarios, siendo el usuario quién define las reglas de funcionamiento. Este trabajo se trata del cuarto ciclo de desarrollo, pues la aplicación ha sido crea y ampliada en otros tres proyectos. En este último desarrollo se han actualizado las tecnologías y herramientas que forman parte de la aplicación. Se ha puesto especial énfasis en el rediseño de la interfaz web, adoptando el uso de las últimas tecnologías web que permiten un funcionamiento dinámico de la misma. Por otro lado se han corregido algunos errores de diseño e introducido el uso de nuevas herramientas para la gestión del proyecto software. Se trata por lo tanto de un ejercicio de refactorización software donde se ha puesto especial atención en conseguir un proyecto actualizado y que utilice metodologías de desarrollo actuales y que posibilite que sea actualizado en un futuro.

Smile detection for user interfaces

[EN]Perceptual User Interfaces (PUIs) aim at facilitating human-computer interaction with the aid of human-like capacities (computer vision, speech recognition, etc.). In PUIs, the human face is a central element, since it conveys not only identity but also other important information, particularly with respect to the user’s mood or emotional state. This paper describes both a face detector and a smile detector for PUIs. Both are suitable for real-time interaction.

Innovative Man Machine Interfaces In Aeronautics

The research activity focused on the study, design and evaluation of innovative human-machine interfaces based on virtual three-dimensional environments. It is based on the brain electrical activities recorded in real time through the electrical impulses emitted by the brain waves of the user. The achieved target is to identify and sort in real time the different brain states and adapt the interface and/or stimuli to the corresponding emotional state of the user. The setup of an experimental facility based on an innovative experimental methodology for “man in the loop" simulation was established. It allowed involving during pilot training in virtually simulated flights, both pilot and flight examiner, in order to compare the subjective evaluations of this latter to the objective measurements of the brain activity of the pilot. This was done recording all the relevant information versus a time-line. Different combinations of emotional intensities obtained, led to an evaluation of the current situational awareness of the user. These results have a great implication in the current training methodology of the pilots, and its use could be extended as a tool that can improve the evaluation of a pilot/crew performance in interacting with the aircraft when performing tasks and procedures, especially in critical situations. This research also resulted in the design of an interface that adapts the control of the machine to the situation awareness of the user. The new concept worked on, aimed at improving the efficiency between a user and the interface, and gaining capacity by reducing the user’s workload and hence improving the system overall safety. This innovative research combining emotions measured through electroencephalography resulted in a human-machine interface that would have three aeronautical related applications: • An evaluation tool during the pilot training; • An input for cockpit environment; • An adaptation tool of the cockpit automation.

Computer simulation methods to study interfacial tensions : from the Ising model to colloidal crystals

In condensed matter systems, the interfacial tension plays a central role for a multitude of phenomena. It is the driving force for nucleation processes, determines the shape and structure of crystalline structures and is important for industrial applications. Despite its importance, the interfacial tension is hard to determine in experiments and also in computer simulations. While for liquid-vapor interfacial tensions there exist sophisticated simulation methods to compute the interfacial tension, current methods for solid-liquid interfaces produce unsatisfactory results.rnrnAs a first approach to this topic, the influence of the interfacial tension on nuclei is studied within the three-dimensional Ising model. This model is well suited because despite its simplicity, one can learn much about nucleation of crystalline nuclei. Below the so-called roughening temperature, nuclei in the Ising model are not spherical anymore but become cubic because of the anisotropy of the interfacial tension. This is similar to crystalline nuclei, which are in general not spherical but more like a convex polyhedron with flat facets on the surface. In this context, the problem of distinguishing between the two bulk phases in the vicinity of the diffuse droplet surface is addressed. A new definition is found which correctly determines the volume of a droplet in a given configuration if compared to the volume predicted by simple macroscopic assumptions.rnrnTo compute the interfacial tension of solid-liquid interfaces, a new Monte Carlo method called ensemble switch method'' is presented which allows to compute the interfacial tension of liquid-vapor interfaces as well as solid-liquid interfaces with great accuracy. In the past, the dependence of the interfacial tension on the finite size and shape of the simulation box has often been neglected although there is a nontrivial dependence on the box dimensions. As a consequence, one needs to systematically increase the box size and extrapolate to infinite volume in order to accurately predict the interfacial tension. Therefore, a thorough finite-size scaling analysis is established in this thesis. Logarithmic corrections to the finite-size scaling are motivated and identified, which are of leading order and therefore must not be neglected. The astounding feature of these logarithmic corrections is that they do not depend at all on the model under consideration. Using the ensemble switch method, the validity of a finite-size scaling ansatz containing the aforementioned logarithmic corrections is carefully tested and confirmed. Combining the finite-size scaling theory with the ensemble switch method, the interfacial tension of several model systems, ranging from the Ising model to colloidal systems, is computed with great accuracy.

Deep learning for computer vision: a comparison between convolutional neural networks and hierarchical temporal memories on object recognition tasks

In recent years, Deep Learning techniques have shown to perform well on a large variety of problems both in Computer Vision and Natural Language Processing, reaching and often surpassing the state of the art on many tasks. The rise of deep learning is also revolutionizing the entire field of Machine Learning and Pattern Recognition pushing forward the concepts of automatic feature extraction and unsupervised learning in general. However, despite the strong success both in science and business, deep learning has its own limitations. It is often questioned if such techniques are only some kind of brute-force statistical approaches and if they can only work in the context of High Performance Computing with tons of data. Another important question is whether they are really biologically inspired, as claimed in certain cases, and if they can scale well in terms of "intelligence". The dissertation is focused on trying to answer these key questions in the context of Computer Vision and, in particular, Object Recognition, a task that has been heavily revolutionized by recent advances in the field. Practically speaking, these answers are based on an exhaustive comparison between two, very different, deep learning techniques on the aforementioned task: Convolutional Neural Network (CNN) and Hierarchical Temporal memory (HTM). They stand for two different approaches and points of view within the big hat of deep learning and are the best choices to understand and point out strengths and weaknesses of each of them. CNN is considered one of the most classic and powerful supervised methods used today in machine learning and pattern recognition, especially in object recognition. CNNs are well received and accepted by the scientific community and are already deployed in large corporation like Google and Facebook for solving face recognition and image auto-tagging problems. HTM, on the other hand, is known as a new emerging paradigm and a new meanly-unsupervised method, that is more biologically inspired. It tries to gain more insights from the computational neuroscience community in order to incorporate concepts like time, context and attention during the learning process which are typical of the human brain. In the end, the thesis is supposed to prove that in certain cases, with a lower quantity of data, HTM can outperform CNN.

Fully Automatic Segmentation of Brain Tumor Images using Support Vector Machine Classiffication in Combination with Hierarchical Conditional Random Field Regularization

Delineating brain tumor boundaries from magnetic resonance images is an essential task for the analysis of brain cancer. We propose a fully automatic method for brain tissue segmentation, which combines Support Vector Machine classification using multispectral intensities and textures with subsequent hierarchical regularization based on Conditional Random Fields. The CRF regularization introduces spatial constraints to the powerful SVM classification, which assumes voxels to be independent from their neighbors. The approach first separates healthy and tumor tissue before both regions are subclassified into cerebrospinal fluid, white matter, gray matter and necrotic, active, edema region respectively in a novel hierarchical way. The hierarchical approach adds robustness and speed by allowing to apply different levels of regularization at different stages. The method is fast and tailored to standard clinical acquisition protocols. It was assessed on 10 multispectral patient datasets with results outperforming previous methods in terms of segmentation detail and computation times.

Correlation properties of spontaneous motor activity in healthy infants: a new computer-assisted method to evaluate neurological maturation

Qualitative assessment of spontaneous motor activity in early infancy is widely used in clinical practice. It enables the description of maturational changes of motor behavior in both healthy infants and infants who are at risk for later neurological impairment. These assessments are, however, time-consuming and are dependent upon professional experience. Therefore, a simple physiological method that describes the complex behavior of spontaneous movements (SMs) in infants would be helpful. In this methodological study, we aimed to determine whether time series of motor acceleration measurements at 40-44 weeks and 50-55 weeks gestational age in healthy infants exhibit fractal-like properties and if this self-affinity of the acceleration signal is sensitive to maturation. Healthy motor state was ensured by General Movement assessment. We assessed statistical persistence in the acceleration time series by calculating the scaling exponent α via detrended fluctuation analysis of the time series. In hand trajectories of SMs in infants we found a mean α value of 1.198 (95 % CI 1.167-1.230) at 40-44 weeks. Alpha changed significantly (p = 0.001) at 50-55 weeks to a mean of 1.102 (1.055-1.149). Complementary multilevel regression analysis confirmed a decreasing trend of α with increasing age. Statistical persistence of fluctuation in hand trajectories of SMs is sensitive to neurological maturation and can be characterized by a simple parameter α in an automated and observer-independent fashion. Future studies including children at risk for neurological impairment should evaluate whether this method could be used as an early clinical screening tool for later neurological compromise.

Application and comparison of classification algorithms for recognition of Alzheimer's disease in electrical brain activity (EEG)

The early detection of subjects with probable Alzheimer's disease (AD) is crucial for effective appliance of treatment strategies. Here we explored the ability of a multitude of linear and non-linear classification algorithms to discriminate between the electroencephalograms (EEGs) of patients with varying degree of AD and their age-matched control subjects. Absolute and relative spectral power, distribution of spectral power, and measures of spatial synchronization were calculated from recordings of resting eyes-closed continuous EEGs of 45 healthy controls, 116 patients with mild AD and 81 patients with moderate AD, recruited in two different centers (Stockholm, New York). The applied classification algorithms were: principal component linear discriminant analysis (PC LDA), partial least squares LDA (PLS LDA), principal component logistic regression (PC LR), partial least squares logistic regression (PLS LR), bagging, random forest, support vector machines (SVM) and feed-forward neural network. Based on 10-fold cross-validation runs it could be demonstrated that even tough modern computer-intensive classification algorithms such as random forests, SVM and neural networks show a slight superiority, more classical classification algorithms performed nearly equally well. Using random forests classification a considerable sensitivity of up to 85% and a specificity of 78%, respectively for the test of even only mild AD patients has been reached, whereas for the comparison of moderate AD vs. controls, using SVM and neural networks, values of 89% and 88% for sensitivity and specificity were achieved. Such a remarkable performance proves the value of these classification algorithms for clinical diagnostics.

High-resolution computer simulation of electrophoretic mobilization in isoelectric focusing

Cationic and anionic electrophoretic mobilization for focusing of hemoglobins (Hb's) in the presence of 100 carrier ampholytes covering a pI range of 6.00-7.98 was studied by computer simulation at a constant current density of 300 A/m(2). Electropherograms that would be produced by whole column imaging and by single detectors placed at different locations along the focusing column are presented. Upon mobilization, peak heights of the Hb zones decrease, but the zones retain a relatively sharp constant profile and are migrating at a constant velocity. A further peak decrease occurs during readjustment at the locations of the original buffer/column interfaces, indicating that detection sensitivity is the lowest at these locations. An anionic carrier ampholyte mobility smaller than that of its cationic species produces a cathodic drift which is smaller than the transport rate used for electrophoretic mobilization. Compared to the case with equal mobilities of carrier ampholyte species, a small increase (decrease) is predicted for the cationic (anionic) mobilization rate within the focusing column. Simulation data suggest that electrophoretic mobilization after focusing and focusing with concurrent electrophoretic mobilization are comparable isotachophoretic processes that occur when there is an uninterrupted flux of an ion through the focusing column. Cathodic drift caused by unequal mobilities of the species of carrier ampholytes, electrophoretic mobilization, and decomposition occurring at the pH gradient edges are related electrophoretic processes.

Assessing Electromyographic Interfaces

Electronic apppliances are increasingly a part of our everyday lives. In particular, mobile devices, with their reduced dimensions with power rivaling desktop computers, have substantially augmented our communication abilities offering instant availability, anywhere, to everyone. These devices have become essential for human communication but also include a more comprehensive tool set to support productivity and leisure applications. However, the many applications commonly available are not adapted to people with special needs. Rather, most popular devices are targeted at teenagers or young adults with excellent eyesight and coordination. What is worse, most of the commonly used assistive control interfaces are not available in a mobile environment where user's position, accommodation and capacities can vary even widely. To try and address people with special needs new approaches and techniques are sorely needed. This paper presents a control interface to allow tetraplegic users to interact with electronic devices. Our method uses myographic information (Electromyography or EMG) collected from residually controlled body areas. User evaluations validate electromyography as a daily wearable interface. In particular our results show that EMG can be used even in mobility contexts.

Microstates in language-related brain potential maps show noun-verb differences

Brain processing of grammatical word class was studied analyzing event-related potential (ERP) brain fields. Normal subjects observed a randomized sequence of single German nouns and verbs on a computer screen, while 20-channel ERP field map series were recorded separately for both word classes. Spatial microstate analysis was applied, based on the observation that series of ERP maps consist of epochs of quasi-stable map landscapes and based on the rationale that different map landscapes must have been generated by different neural generators and thus suggest different brain functions. Space-oriented segmentation of the mean map series identified nine successive, different functional microstates, i.e., steps of brain information processing characterized by quasi-stable map landscapes. In the microstate from 116 to 172 msec, noun-related maps differed significantly from verb-related maps along the left–right axis. The results indicate that different neural populations represent different grammatical word classes in language processing, in agreement with clinical observations. This word class differentiation as revealed by the spatial–temporal organization of neural activity occurred at a time after word input compatible with speed of reading.

Hierarchical segmentation-assisted multimodal registration for MR brain images

Information theory-based metric such as mutual information (MI) is widely used as similarity measurement for multimodal registration. Nevertheless, this metric may lead to matching ambiguity for non-rigid registration. Moreover, maximization of MI alone does not necessarily produce an optimal solution. In this paper, we propose a segmentation-assisted similarity metric based on point-wise mutual information (PMI). This similarity metric, termed SPMI, enhances the registration accuracy by considering tissue classification probabilities as prior information, which is generated from an expectation maximization (EM) algorithm. Diffeomorphic demons is then adopted as the registration model and is optimized in a hierarchical framework (H-SPMI) based on different levels of anatomical structure as prior knowledge. The proposed method is evaluated using Brainweb synthetic data and clinical fMRI images. Both qualitative and quantitative assessment were performed as well as a sensitivity analysis to the segmentation error. Compared to the pure intensity-based approaches which only maximize mutual information, we show that the proposed algorithm provides significantly better accuracy on both synthetic and clinical data.

Integrated Segmentation of Brain Tumor Images for Radiotherapy and Neurosurgery

Image-based modeling of tumor growth combines methods from cancer simulation and medical imaging. In this context, we present a novel approach to adapt a healthy brain atlas to MR images of tumor patients. In order to establish correspondence between a healthy atlas and a pathologic patient image, tumor growth modeling in combination with registration algorithms is employed. In a first step, the tumor is grown in the atlas based on a new multi-scale, multi-physics model including growth simulation from the cellular level up to the biomechanical level, accounting for cell proliferation and tissue deformations. Large-scale deformations are handled with an Eulerian approach for finite element computations, which can operate directly on the image voxel mesh. Subsequently, dense correspondence between the modified atlas and patient image is established using nonrigid registration. The method offers opportunities in atlasbased segmentation of tumor-bearing brain images as well as for improved patient-specific simulation and prognosis of tumor progression.

Computer simulation of glioma growth and morphology.

Despite major advances in the study of glioma, the quantitative links between intra-tumor molecular/cellular properties, clinically observable properties such as morphology, and critical tumor behaviors such as growth and invasiveness remain unclear, hampering more effective coupling of tumor physical characteristics with implications for prognosis and therapy. Although molecular biology, histopathology, and radiological imaging are employed in this endeavor, studies are severely challenged by the multitude of different physical scales involved in tumor growth, i.e., from molecular nanoscale to cell microscale and finally to tissue centimeter scale. Consequently, it is often difficult to determine the underlying dynamics across dimensions. New techniques are needed to tackle these issues. Here, we address this multi-scalar problem by employing a novel predictive three-dimensional mathematical and computational model based on first-principle equations (conservation laws of physics) that describe mathematically the diffusion of cell substrates and other processes determining tumor mass growth and invasion. The model uses conserved variables to represent known determinants of glioma behavior, e.g., cell density and oxygen concentration, as well as biological functional relationships and parameters linking phenomena at different scales whose specific forms and values are hypothesized and calculated based on in vitro and in vivo experiments and from histopathology of tissue specimens from human gliomas. This model enables correlation of glioma morphology to tumor growth by quantifying interdependence of tumor mass on the microenvironment (e.g., hypoxia, tissue disruption) and on the cellular phenotypes (e.g., mitosis and apoptosis rates, cell adhesion strength). Once functional relationships between variables and associated parameter values have been informed, e.g., from histopathology or intra-operative analysis, this model can be used for disease diagnosis/prognosis, hypothesis testing, and to guide surgery and therapy. In particular, this tool identifies and quantifies the effects of vascularization and other cell-scale glioma morphological characteristics as predictors of tumor-scale growth and invasion.