Contributions to unsupervised and supervised learning with applications in digital image processing

311 p. : il.

Structural Integrity of the Contralesional Hemisphere Predicts Cognitive Impairment in Ischemic Stroke at Three Months

After stroke, white matter integrity can be affected both locally and distally to the primary lesion location. It has been shown that tract disruption in mirror's regions of the contralateral hemisphere is associated with degree of functional impairment. Fourteen patients suffering right hemispheric focal stroke (S) and eighteen healthy controls (HC) underwent Diffusion Weighted Imaging (DWI) and neuropsychological assessment. The stroke patient group was divided into poor (SP; n = 8) and good (SG; n = 6) cognitive recovery groups according to their cognitive improvement from the acute phase (72 hours after stroke) to the subacute phase (3 months post-stroke). Whole-brain DWI data analysis was performed by computing Diffusion Tensor Imaging (DTI) followed by Tract Based Spatial Statistics (TBSS). Assessment of effects was obtained computing the correlation of the projections on TBSS skeleton of Fractional Anisotropy (FA) and Radial Diffusivity (RD) with cognitive test results. Significant decrease of FA was found only in right brain anatomical areas for the S group when compared to the HC group. Analyzed separately, stroke patients with poor cognitive recovery showed additional significant FA decrease in several left hemisphere regions; whereas SG patients showed significant decrease only in the left genu of corpus callosum when compared to the HC. For the SG group, whole brain analysis revealed significant correlation between the performance in the Semantic Fluency test and the FA in the right hemisphere as well as between the performance in the Grooved Pegboard Test (GPT) and theTrail Making Test-part A and the FA in the left hemisphere. For the SP group, correlation analysis revealed significant correlation between the performance in the GPT and the FA in the right hemisphere. Palabras clave

Skeleton based visual pattern recognition. Applications to tabletop interaction

211 p. :il.

Contributions to the analysis and segmentation of remote sensing hyperspectral images

142 p.

Experimentos de coordinación multi-robot con quadcopteros

El objetivo principal de este proyecto es crear un sistema capaz de controlar varios UAVs y hacer experimentos con ellos de manera coordinada. El UAV utilizado será la plataforma robótica aérea AR.Drone. Estos drones son cuadricópteros con sus cuatro servomotores eléctricos que permiten un control muy robusto de sus maniobras en el aire. El sistema completo estará compuesto por varios drones y un controlador, que en este caso será un ordenador. A partir de la creación de la red compuesta por los drones y el controlador, se detallarán los programas que se han utilizado para controlar los drones, ya sea en los vuelos autónomos o vuelos controlados. El objetivo de estos vuelos será transportar objetos lineales, como pueden ser mangueras o cuerdas, mediante el vuelo coordinado de los drones dotados de sistemas de jación para que los puedan transportar. Esto es, gracias a un sistema de sujeción que le añadiremos a cada drone probaremos el transporte de varios tipos de cuerdas por dos o más drones a la vez guiados por un solo controlador. Los programas creados tendrán el objetivo de corregir el vuelo e intentar conseguir la estabilidad necesaria para que los drones puedan transportar las cuerdas o mangueras sin perder el control debido a su peso o algún efecto producido por los otros drones, como corrientes de aire inducidas por los rotores. Este proyecto tiene dos partes importantes: La primera es la creación del sistema que nos permite transportar las cuerdas y la segunda es observar y analizar el comportamiento del sistema durante diferentes experimentos. A la hora de explicar las pruebas experimentales se detallará la situación del sistema con el número de drones y el objeto a transportar. Además, se resumirán los resultados de los experimentos re ejados en valores de parámetros de vuelo recibidos de los drones. Para acabar, se detallarán las diferentes conclusiones a las que se ha llegado mediante los experimentos y que nos han servido también para escribir las lineas futuras de trabajo que se detallan al nal.

Instalación de sistema contra incendios de un almacen de calzado.

Diseño, realización de planos y cálculos necesarios para la instalación de un sistema de protección contra incendios en un almacen de calzado.

Energy demands of diverse spiking cells from the neocortex, hippocampus, and thalamus

It has long been known that neurons in the brain are not physiologically homogeneous. In response to current stimulus, they can fire several distinct patterns of action potentials that are associated with different physiological classes ranging from regular-spiking cells, fast-spiking cells, intrinsically bursting cells, and low-threshold cells. In this work we show that the high degree of variability in firing characteristics of action potentials among these cells is accompanied with a significant variability in the energy demands required to restore the concentration gradients after an action potential. The values of the metabolic energy were calculated for a wide range of cell temperatures and stimulus intensities following two different approaches. The first one is based on the amount of Na+ load crossing the membrane during a single action potential, while the second one focuses on the electrochemical energy functions deduced from the dynamics of the computational neuron models. The results show that the thalamocortical relay neuron is the most energy-efficient cell consuming between 7 and 18 nJ/cm(2) for each spike generated, while both the regular and fast spiking cells from somatosensory cortex and the intrinsically-bursting cell from a cat visual cortex are the least energy-efficient, and can consume up to 100 nJ/cm(2) per spike. The lowest values of these energy demands were achieved at higher temperatures and high external stimuli.

Structural Integrity of the Contralesional Hemisphere Predicts Cognitive Impairment in Ischemic Stroke at Three Months

After stroke, white matter integrity can be affected both locally and distally to the primary lesion location. It has been shown that tract disruption in mirror's regions of the contralateral hemisphere is associated with degree of functional impairment. Fourteen patients suffering right hemispheric focal stroke (S) and eighteen healthy controls (HC) underwent Diffusion Weighted Imaging (DWI) and neuropsychological assessment. The stroke patient group was divided into poor (SP; n = 8) and good (SG; n = 6) cognitive recovery groups according to their cognitive improvement from the acute phase (72 hours after stroke) to the subacute phase (3 months post-stroke). Whole-brain DWI data analysis was performed by computing Diffusion Tensor Imaging (DTI) followed by Tract Based Spatial Statistics (TBSS). Assessment of effects was obtained computing the correlation of the projections on TBSS skeleton of Fractional Anisotropy (FA) and Radial Diffusivity (RD) with cognitive test results. Significant decrease of FA was found only in right brain anatomical areas for the S group when compared to the HC group. Analyzed separately, stroke patients with poor cognitive recovery showed additional significant FA decrease in several left hemisphere regions; whereas SG patients showed significant decrease only in the left genu of corpus callosum when compared to the HC. For the SG group, whole brain analysis revealed significant correlation between the performance in the Semantic Fluency test and the FA in the right hemisphere as well as between the performance in the Grooved Pegboard Test (GPT) and theTrail Making Test-part A and the FA in the left hemisphere. For the SP group, correlation analysis revealed significant correlation between the performance in the GPT and the FA in the right hemisphere.

Contributions to Ensemble Classifiers with Image Analysis Applications

134 p.

Learning Multirobot Hose Transportation and Deployment by Distributed Round-Robin Q-Learning

Multi-Agent Reinforcement Learning (MARL) algorithms face two main difficulties: the curse of dimensionality, and environment non-stationarity due to the independent learning processes carried out by the agents concurrently. In this paper we formalize and prove the convergence of a Distributed Round Robin Q-learning (D-RR-QL) algorithm for cooperative systems. The computational complexity of this algorithm increases linearly with the number of agents. Moreover, it eliminates environment non sta tionarity by carrying a round-robin scheduling of the action selection and execution. That this learning scheme allows the implementation of Modular State-Action Vetoes (MSAV) in cooperative multi-agent systems, which speeds up learning convergence in over-constrained systems by vetoing state-action pairs which lead to undesired termination states (UTS) in the relevant state-action subspace. Each agent's local state-action value function learning is an independent process, including the MSAV policies. Coordination of locally optimal policies to obtain the global optimal joint policy is achieved by a greedy selection procedure using message passing. We show that D-RR-QL improves over state-of-the-art approaches, such as Distributed Q-Learning, Team Q-Learning and Coordinated Reinforcement Learning in a paradigmatic Linked Multi-Component Robotic System (L-MCRS) control problem: the hose transportation task. L-MCRS are over-constrained systems with many UTS induced by the interaction of the passive linking element and the active mobile robots.