Multi-resolution model-based traffic sign detection and tracking

In this paper we propose an innovative approach to tackle the problem of traffic sign detection using a computer vision algorithm and taking into account real-time operation constraints, trying to establish intelligent strategies to simplify as much as possible the algorithm complexity and to speed up the process. Firstly, a set of candidates is generated according to a color segmentation stage, followed by a region analysis strategy, where spatial characteristic of previously detected objects are taken into account. Finally, temporal coherence is introduced by means of a tracking scheme, performed using a Kalman filter for each potential candidate. Taking into consideration time constraints, efficiency is achieved two-fold: on the one side, a multi-resolution strategy is adopted for segmentation, where global operation will be applied only to low-resolution images, increasing the resolution to the maximum only when a potential road sign is being tracked. On the other side, we take advantage of the expected spacing between traffic signs. Namely, the tracking of objects of interest allows to generate inhibition areas, which are those ones where no new traffic signs are expected to appear due to the existence of a TS in the neighborhood. The proposed solution has been tested with real sequences in both urban areas and highways, and proved to achieve higher computational efficiency, especially as a result of the multi-resolution approach.

HOG-Like gradient-based descriptor for visual vehicle detection

One of the main challenges for intelligent vehicles is the capability of detecting other vehicles in their environment, which constitute the main source of accidents. Specifically, many methods have been proposed in the literature for video-based vehicle detection. Most of them perform supervised classification using some appearance-related feature, in particular, symmetry has been extensively utilized. However, an in-depth analysis of the classification power of this feature is missing. As a first contribution of this paper, a thorough study of the classification performance of symmetry is presented within a Bayesian decision framework. This study reveals that the performance of symmetry-based classification is very limited. Therefore, as a second contribution, a new gradient-based descriptor is proposed for vehicle detection. This descriptor exploits the known rectangular structure of vehicle rears within a Histogram of Gradients (HOG)-based framework. Experiments show that the proposed descriptor outperforms largely symmetry as a feature for vehicle verification, achieving classification rates over 90%.

Moving object detection strategy for augmented-reality applications in a GPGPU by using CUDA

A spatial-color-based non-parametric background-foreground modeling strategy in a GPGPU by using CUDA is proposed. This strategy is suitable for augmented-reality applications, providing real-time high-quality results in a great variety of scenarios.

Temporal segmentation tool for high-quality real time video editing software

The increasing use of video editing software requires faster and more efficient editing tools. As a first step, these tools perform a temporal segmentation in shots that allows a later building of indexes describing the video content. Here, we propose a novel real-time high-quality shot detection strategy, suitable for the last generation of video editing software requiring both low computational cost and high quality results. While abrupt transitions are detected through a very fast pixel-based analysis, gradual transitions are obtained from an efficient edge-based analysis. Both analyses are reinforced with a motion analysis that helps to detect and discard false detections. This motion analysis is carried out exclusively over a reduced set of candidate transitions, thus maintaining the computational requirements demanded by new applications to fulfill user needs.

Providing 3D video services: the challenge from 2D to 3DTV quality of experience

Recently, three-dimensional (3D) video has decisively burst onto the entertainment industry scene, and has arrived in households even before the standardization process has been completed. 3D television (3DTV) adoption and deployment can be seen as a major leap in television history, similar to previous transitions from black and white (B&W) to color, from analog to digital television (TV), and from standard definition to high definition. In this paper, we analyze current 3D video technology trends in order to define a taxonomy of the availability and possible introduction of 3D-based services. We also propose an audiovisual network services architecture which provides a smooth transition from two-dimensional (2D) to 3DTV in an Internet Protocol (IP)-based scenario. Based on subjective assessment tests, we also analyze those factors which will influence the quality of experience in those 3D video services, focusing on effects of both coding and transmission errors. In addition, examples of the application of the architecture and results of assessment tests are provided.

Moving object detection for real-time augmented reality applications in a GPGPU

The last generation of consumer electronic devices is endowed with Augmented Reality (AR) tools. These tools require moving object detection strategies, which should be fast and efficient, to carry out higher level object analysis tasks. We propose a lightweight spatio-temporal-based non-parametric background-foreground modeling strategy in a General Purpose Graphics Processing Unit (GPGPU), which provides real-time high-quality results in a great variety of scenarios and is suitable for AR applications.

The Role of Knowledge Modeling Techniques in Software Development: A General Approach Based on a Knowledge Management Tool

The aim of the paper is to discuss the use of knowledge models to formulate general applications. First, the paper presents the recent evolution of the software field where increasing attention is paid to conceptual modeling. Then, the current state of knowledge modeling techniques is described where increased reliability is available through the modern knowledge acquisition techniques and supporting tools. The KSM (Knowledge Structure Manager) tool is described next. First, the concept of knowledge area is introduced as a building block where methods to perform a collection of tasks are included together with the bodies of knowledge providing the basic methods to perform the basic tasks. Then, the CONCEL language to define vocabularies of domains and the LINK language for methods formulation are introduced. Finally, the object oriented implementation of a knowledge area is described and a general methodology for application design and maintenance supported by KSM is proposed. To illustrate the concepts and methods, an example of system for intelligent traffic management in a road network is described. This example is followed by a proposal of generalization for reuse of the resulting architecture. Finally, some concluding comments are proposed about the feasibility of using the knowledge modeling tools and methods for general application design.

High Efficiency Envelope Amplifier based on a Ripple Cancellation Buck Converter. Design, Optimization and Integration in an EER RFPA

Durante los ��ltimos a��os la tendencia en el sector de las telecomunicaciones ha sido un aumento y diversificaci��n en la transmisi��n de voz, video y fundamentalmente de datos. Para conseguir alcanzar las tasas de transmisi��n requeridas, los nuevos est��ndares de comunicaciones requieren un mayor ancho de banda y tienen un mayor factor de pico, lo cual influye en el bajo rendimiento del amplificador de radiofrecuencia (RFPA). Otro factor que ha influido en el bajo rendimiento es el dise��o del amplificador de radiofrecuencia. Tradicionalmente se han utilizado amplificadores lineales por su buen funcionamiento. Sin embargo, debido al elevado factor de pico de las se��ales transmitidas, el rendimiento de este tipo de amplificadores es bajo. El bajo rendimiento del sistema conlleva desventajas adicionales como el aumento del coste y del tama��o del sistema de refrigeraci��n, como en el caso de una estaci��n base, o como la reducci��n del tiempo de uso y un mayor calentamiento del equipo para sistemas port��tiles alimentados con bater��as. Debido a estos factores, se han desarrollado durante las ��ltimas d��cadas varias soluciones para aumentar el rendimiento del RFPA como la t��cnica de Outphasing, combinadores de potencia o la t��cnica de Doherty. Estas soluciones mejoran las prestaciones del RFPA y en alg��n caso han sido ampliamente utilizados comercialmente como la t��cnica de Doherty, que alcanza rendimientos hasta del 50% para el sistema completo para anchos de banda de hasta 20MHz. Pese a las mejoras obtenidas con estas soluciones, los mayores rendimientos del sistema se obtienen para soluciones basadas en la modulaci��n de la tensi��n de alimentaci��n del amplificador de potencia como ���Envelope Tracking��� o ���EER���. La t��cnica de seguimiento de envolvente o ���Envelope Tracking��� est�� basada en la modulaci��n de la tensi��n de alimentaci��n de un amplificador lineal de potencia para obtener una mejora en el rendimiento en el sistema comparado a una soluci��n con una tensi��n de alimentaci��n constante. Para la implementaci��n de esta t��cnica se necesita una etapa adicional, el amplificador de envolvente, que a��ade complejidad al amplificador de radiofrecuencia. En un amplificador dise��ado con esta t��cnica, se aumentan las p��rdidas debido a la etapa adicional que supone el amplificador de envolvente pero a su vez disminuyen las p��rdidas en el amplificador de potencia. Si el dise��o se optimiza adecuadamente, puede conseguirse un aumento global en el rendimiento del sistema superior al conseguido con las t��cnicas mencionadas anteriormente. Esta t��cnica presenta ventajas en el dise��o del amplificador de envolvente, ya que el ancho de banda requerido puede ser menor que el ancho de banda de la se��al de envolvente si se optimiza adecuadamente el dise��o. Adicionalmente, debido a que la sincronizaci��n entre la se��al de envolvente y de fase no tiene que ser perfecta, el proceso de integraci��n conlleva ciertas ventajas respecto a otras t��cnicas como EER. La t��cnica de eliminaci��n y restauraci��n de envolvente, llamada EER o t��cnica de Kahn est�� basada en modulaci��n simult��nea de la envolvente y la fase de la se��al usando un amplificador de potencia conmutado, no lineal y que permite obtener un elevado rendimiento. Esta soluci��n fue propuesta en el a��o 1952, pero no ha sido implementada con ��xito durante muchos a��os debido a los exigentes requerimientos en cuanto a la sincronizaci��n entre fase y envolvente, a las t��cnicas de control y de correcci��n de los errores y no linealidades de cada una de las etapas as�� como de los equipos para poder implementar estas t��cnicas, que tienen unos requerimientos exigentes en capacidad de c��lculo y procesamiento. Dentro del dise��o de un RFPA, el amplificador de envolvente tiene una gran importancia debido a su influencia en el rendimiento y ancho de banda del sistema completo. Adicionalmente, la linealidad y la calidad de la se��al de transmitida deben ser elevados para poder cumplir con los diferentes est��ndares de telecomunicaciones. Esta tesis se centra en el amplificador de envolvente y el objetivo principal es el desarrollo de soluciones que permitan el aumento del rendimiento total del sistema a la vez que satisfagan los requerimientos de ancho de banda, calidad de la se��al transmitida y de linealidad. Debido al elevado rendimiento que potencialmente puede alcanzarse con la t��cnica de EER, esta t��cnica ha sido objeto de an��lisis y en el estado del arte pueden encontrarse numerosas referencias que analizan el dise��o y proponen diversas implementaciones. En una clasificaci��n de alto nivel, podemos agrupar las soluciones propuestas del amplificador de envolvente seg��n est��n compuestas de una o m��ltiples etapas. Las soluciones para el amplificador de envolvente en una configuraci��n multietapa se basan en la combinaci��n de un convertidor conmutado, de elevado rendimiento con un regulador lineal, de alto ancho de banda, en una combinaci��n serie o paralelo. Estas soluciones, debido a la combinaci��n de las caracter��sticas de ambas etapas, proporcionan un buen compromiso entre rendimiento y buen funcionamiento del amplificador de RF. Por otro lado, la complejidad del sistema aumenta debido al mayor n��mero de componentes y de se��ales de control necesarias y el aumento de rendimiento que se consigue con estas soluciones es limitado. Una configuraci��n en una etapa tiene las ventajas de una mayor simplicidad, pero debido al elevado ancho de banda necesario, la frecuencia de conmutaci��n debe aumentarse en gran medida. Esto implicar�� un bajo rendimiento y un peor funcionamiento del amplificador de envolvente. En el estado del arte pueden encontrarse diversas soluciones para un amplificador de envolvente en una etapa, como aumentar la frecuencia de conmutaci��n y realizar la implementaci��n en un circuito integrado, que tendr�� mejor funcionamiento a altas frecuencias o utilizar t��cnicas topol��gicas y/o filtros de orden elevado, que permiten una reducci��n de la frecuencia de conmutaci��n. En esta tesis se propone de manera original el uso de la t��cnica de cancelaci��n de rizado, aplicado al convertidor reductor s��ncrono, para reducir la frecuencia de conmutaci��n comparado con dise��o equivalente del convertidor reductor convencional. Adicionalmente se han desarrollado dos variantes topol��gicas basadas en esta soluci��n para aumentar la robustez y las prestaciones de la misma. Otro punto de inter��s en el dise��o de un RFPA es la dificultad de poder estimar la influencia de los par��metros de dise��o del amplificador de envolvente en el amplificador final integrado. En esta tesis se ha abordado este problema y se ha desarrollado una herramienta de dise��o que permite obtener las principales figuras de m��rito del amplificador integrado para la t��cnica de EER a partir del dise��o del amplificador de envolvente. Mediante el uso de esta herramienta pueden validarse el efecto del ancho de banda, el rizado de tensi��n de salida o las no linealidades del dise��o del amplificador de envolvente para varias modulaciones digitales. Las principales contribuciones originales de esta tesis son las siguientes: ������La aplicaci��n de la t��cnica de cancelaci��n de rizado a un convertidor reductor s��ncrono para un amplificador de envolvente de alto rendimiento para un RFPA linealizado mediante la t��cnica de EER. ������Una reducci��n del 66% en la frecuencia de conmutaci��n, comparado con el reductor convencional equivalente. Esta reducci��n se ha validado experimentalmente obteni��ndose una mejora en el rendimiento de entre el 12.4% y el 16% para las especificaciones de este trabajo. ������La topolog��a y el dise��o del convertidor reductor con dos redes de cancelaci��n de rizado en cascada para mejorar el funcionamiento y robustez de la soluci��n con una red de cancelaci��n. ������La combinaci��n de un convertidor redactor multifase con la t��cnica de cancelaci��n de rizado para obtener una topolog��a que proporciona una reducci��n del cociente entre frecuencia de conmutaci��n y ancho de banda de la se��al. ������El proceso de optimizaci��n del control del amplificador de envolvente en lazo cerrado para mejorar el funcionamiento respecto a la soluci��n en lazo abierto del convertidor reductor con red de cancelaci��n de rizado. ������Una herramienta de simulaci��n para optimizar el proceso de dise��o del amplificador de envolvente mediante la estimaci��n de las figuras de m��rito del RFPA, implementado mediante EER, basada en el dise��o del amplificador de envolvente. ������La integraci��n y caracterizaci��n del amplificador de envolvente basado en un convertidor reductor con red de cancelaci��n de rizado en el transmisor de radiofrecuencia completo consiguiendo un elevado rendimiento, entre 57% y 70.6% para potencias de salida de 14.4W y 40.7W respectivamente. Esta tesis se divide en seis cap��tulos. El primer cap��tulo aborda la introducci��n enfocada en la aplicaci��n, los amplificadores de potencia de radiofrecuencia, as�� como los principales problemas, retos y soluciones existentes. En el cap��tulo dos se desarrolla el estado del arte de amplificadores de potencia de RF, describi��ndose las principales t��cnicas de dise��o, las causas de no linealidad y las t��cnicas de optimizaci��n. El cap��tulo tres est�� centrado en las soluciones propuestas para el amplificador de envolvente. El modo de control se ha abordado en este cap��tulo y se ha presentado una optimizaci��n del dise��o en lazo cerrado para el convertidor reductor convencional y para el convertidor reductor con red de cancelaci��n de rizado. El cap��tulo cuatro se centra en el proceso de dise��o del amplificador de envolvente. Se ha desarrollado una herramienta de dise��o para evaluar la influencia del amplificador de envolvente en las figuras de m��rito del RFPA. En el cap��tulo cinco se presenta el proceso de integraci��n realizado y las pruebas realizadas para las diversas modulaciones, as�� como la completa caracterizaci��n y an��lisis del amplificador de RF. El cap��tulo seis describe las principales conclusiones de la tesis y las l��neas futuras. ABSTRACT The trend in the telecommunications sector during the last years follow a high increase in the transmission rate of voice, video and mainly in data. To achieve the required levels of data rates, the new modulation standards demand higher bandwidths and have a higher peak to average power ratio (PAPR). These specifications have a direct impact in the low efficiency of the RFPA. An additional factor for the low efficiency of the RFPA is in the power amplifier design. Traditionally, linear classes have been used for the implementation of the power amplifier as they comply with the technical requirements. However, they have a low efficiency, especially in the operating range of signals with a high PAPR. The low efficiency of the transmitter has additional disadvantages as an increase in the cost and size as the cooling system needs to be increased for a base station and a temperature increase and a lower use time for portable devices. Several solutions have been proposed in the state of the art to improve the efficiency of the transmitter as Outphasing, power combiners or Doherty technique. However, the highest potential of efficiency improvement can be obtained using a modulated power supply for the power amplifier, as in the Envelope Tracking and EER techniques. The Envelope Tracking technique is based on the modulation of the power supply of a linear power amplifier to improve the overall efficiency compared to a fixed voltage supply. In the implementation of this technique an additional stage is needed, the envelope amplifier, that will increase the complexity of the RFPA. However, the efficiency of the linear power amplifier will increase and, if designed properly, the RFPA efficiency will be improved. The advantages of this technique are that the envelope amplifier design does not require such a high bandwidth as the envelope signal and that in the integration process a perfect synchronization between envelope and phase is not required. The Envelope Elimination and Restoration (EER) technique, known also as Kahn���s technique, is based on the simultaneous modulation of envelope and phase using a high efficiency switched power amplifier. This solution has the highest potential in terms of the efficiency improvement but also has the most challenging specifications. This solution, proposed in 1952, has not been successfully implemented until the last two decades due to the high demanding requirements for each of the stages as well as for the highly demanding processing and computation capabilities needed. At the system level, a very precise synchronization is required between the envelope and phase paths to avoid a linearity decrease of the system. Several techniques are used to compensate the non-linear effects in amplitude and phase and to improve the rejection of the out of band noise as predistortion, feedback and feed-forward. In order to obtain a high bandwidth and efficient RFPA using either ET or EER, the envelope amplifier stage will have a critical importance. The requirements for this stage are very demanding in terms of bandwidth, linearity and quality of the transmitted signal. Additionally the efficiency should be as high as possible, as the envelope amplifier has a direct impact in the efficiency of the overall system. This thesis is focused on the envelope amplifier stage and the main objective will be the development of high efficiency envelope amplifier solutions that comply with the requirements of the RFPA application. The design and optimization of an envelope amplifier for a RFPA application is a highly referenced research topic, and many solutions that address the envelope amplifier and the RFPA design and optimization can be found in the state of the art. From a high level classification, multiple and single stage envelope amplifiers can be identified. Envelope amplifiers for EER based on multiple stage architecture combine a linear assisted stage and a switched-mode stage, either in a series or parallel configuration, to achieve a very high performance RFPA. However, the complexity of the system increases and the efficiency improvement is limited. A single-stage envelope amplifier has the advantage of a lower complexity but in order to achieve the required bandwidth the switching frequency has to be highly increased, and therefore the performance and the efficiency are degraded. Several techniques are used to overcome this limitation, as the design of integrated circuits that are capable of switching at very high rates or the use of topological solutions, high order filters or a combination of both to reduce the switching frequency requirements. In this thesis it is originally proposed the use of the ripple cancellation technique, applied to a synchronous buck converter, to reduce the switching frequency requirements compared to a conventional buck converter for an envelope amplifier application. Three original proposals for the envelope amplifier stage, based on the ripple cancellation technique, are presented and one of the solutions has been experimentally validated and integrated in the complete amplifier, showing a high total efficiency increase compared to other solutions of the state of the art. Additionally, the proposed envelope amplifier has been integrated in the complete RFPA achieving a high total efficiency. The design process optimization has also been analyzed in this thesis. Due to the different figures of merit between the envelope amplifier and the complete RFPA it is very difficult to obtain an optimized design for the envelope amplifier. To reduce the design uncertainties, a design tool has been developed to provide an estimation of the RFPA figures of merit based on the design of the envelope amplifier. The main contributions of this thesis are: ������The application of the ripple cancellation technique to a synchronous buck converter for an envelope amplifier application to achieve a high efficiency and high bandwidth EER RFPA. ������A 66% reduction of the switching frequency, validated experimentally, compared to the equivalent conventional buck converter. This reduction has been reflected in an improvement in the efficiency between 12.4% and 16%, validated for the specifications of this work. ������The synchronous buck converter with two cascaded ripple cancellation networks (RCNs) topology and design to improve the robustness and the performance of the envelope amplifier. ������The combination of a phase-shifted multi-phase buck converter with the ripple cancellation technique to improve the envelope amplifier switching frequency to signal bandwidth ratio. ������The optimization of the control loop of an envelope amplifier to improve the performance of the open loop design for the conventional and ripple cancellation buck converter. ������A simulation tool to optimize the envelope amplifier design process. Using the envelope amplifier design as the input data, the main figures of merit of the complete RFPA for an EER application are obtained for several digital modulations. ������The successful integration of the envelope amplifier based on a RCN buck converter in the complete RFPA obtaining a high efficiency integrated amplifier. The efficiency obtained is between 57% and 70.6% for an output power of 14.4W and 40.7W respectively. The main figures of merit for the different modulations have been characterized and analyzed. This thesis is organized in six chapters. In Chapter 1 is provided an introduction of the RFPA application, where the main problems, challenges and solutions are described. In Chapter 2 the technical background for radiofrequency power amplifiers (RF) is presented. The main techniques to implement an RFPA are described and analyzed. The state of the art techniques to improve performance of the RFPA are identified as well as the main sources of no-linearities for the RFPA. Chapter 3 is focused on the envelope amplifier stage. The three different solutions proposed originally in this thesis for the envelope amplifier are presented and analyzed. The control stage design is analyzed and an optimization is proposed both for the conventional and the RCN buck converter. Chapter 4 is focused in the design and optimization process of the envelope amplifier and a design tool to evaluate the envelope amplifier design impact in the RFPA is presented. Chapter 5 shows the integration process of the complete amplifier. Chapter 6 addresses the main conclusions of the thesis and the future work.

A knowledge based application for industrialization design

After being designed, a product has to be manufactured, which means converting concepts and information into a real, physical object. This requires a big amount of resources and a careful planning. The product manufacturing must be designed too, and that is called Industrialization Design. An accepted methodology for this activity is starting defining simple structures and then progressively increasing the detail degree of the manufacturing solution. The impact of decisions taken at first stages of Industrialization Design is remarkable, and software tools to assist designers are required. In this paper a Knowledge Based Application prototype for the Industrialization Design is presented. The application is implemented within the environment CATIA V5/DELMIA. A case study with a simple Product from aerospace sector illustrates the prototype development.

Correction of the Colour Variations under Uncontrolled Lighting Conditions

Esta tesis trata sobre m��todos de correcci��n que compensan la variaci��n de las condiciones de iluminaci��n en aplicaciones de imagen y video a color. Estas variaciones hacen que a menudo fallen aquellos algoritmos de visi��n artificial que utilizan caracter��sticas de color para describir los objetos. Se formulan tres preguntas de investigaci��n que definen el marco de trabajo de esta tesis. La primera cuesti��n aborda las similitudes que se dan entre las im��genes de superficies adyacentes en relaci��n a su comportamiento fotom��trico. En base al an��lisis del modelo de formaci��n de im��genes en situaciones din��micas, esta tesis propone un modelo capaz de predecir las variaciones de color de la regi��n de una determinada imagen a partir de las variaciones de las regiones colindantes. Dicho modelo se denomina Quotient Relational Model of Regions. Este modelo es v��lido cuando: las fuentes de luz iluminan todas las superficies inclu��das en ��l; estas superficies est��n pr��ximas entre s�� y tienen orientaciones similares; y cuando son en su mayor��a lambertianas. Bajo ciertas circunstancias, la respuesta fotom��trica de una regi��n se puede relacionar con el resto mediante una combinaci��n lineal. No se ha podido encontrar en la literatura cient��fica ning��n trabajo previo que proponga este tipo de modelo relacional. La segunda cuesti��n va un paso m��s all�� y se pregunta si estas similitudes se pueden utilizar para corregir variaciones fotom��tricas desconocidas en una regi��n tambi��n desconocida, a partir de regiones conocidas adyacentes. Para ello, se propone un m��todo llamado Linear Correction Mapping capaz de dar una respuesta afirmativa a esta cuesti��n bajo las circunstancias caracterizadas previamente. Para calcular los par��metros del modelo se requiere una etapa de entrenamiento previo. El m��todo, que inicialmente funciona para una sola c��mara, se ampl��a para funcionar en arquitecturas con varias c��maras sin solape entre sus campos visuales. Para ello, tan solo se necesitan varias muestras de im��genes del mismo objeto capturadas por todas las c��maras. Adem��s, este m��todo tiene en cuenta tanto las variaciones de iluminaci��n, como los cambios en los par��metros de exposici��n de las c��maras. Todos los m��todos de correcci��n de imagen fallan cuando la imagen del objeto que tiene que ser corregido est�� sobreexpuesta o cuando su relaci��n se��al a ruido es muy baja. As��, la tercera cuesti��n se refiere a si se puede establecer un proceso de control de la adquisici��n que permita obtener una exposici��n ��ptima cuando las condiciones de iluminaci��n no est��n controladas. De este modo, se propone un m��todo denominado Camera Exposure Control capaz de mantener una exposici��n adecuada siempre y cuando las variaciones de iluminaci��n puedan recogerse dentro del margen din��mico de la c��mara. Los m��todos propuestos se evaluaron individualmente. La metodolog��a llevada a cabo en los experimentos consisti�� en, primero, seleccionar algunos escenarios que cubrieran situaciones representativas donde los m��todos fueran v��lidos te��ricamente. El Linear Correction Mapping fue validado en tres aplicaciones de re-identificaci��n de objetos (veh��culos, caras y personas) que utilizaban como caracter��siticas la distribuci��n de color de ��stos. Por otra parte, el Camera Exposure Control se prob�� en un parking al aire libre. Adem��s de esto, se definieron varios indicadores que permitieron comparar objetivamente los resultados de los m��todos propuestos con otros m��todos relevantes de correcci��n y auto exposici��n referidos en el estado del arte. Los resultados de la evaluaci��n demostraron que los m��todos propuestos mejoran los m��todos comparados en la mayor��a de las situaciones. Bas��ndose en los resultados obtenidos, se puede decir que las respuestas a las preguntas de investigaci��n planteadas son afirmativas, aunque en circunstancias limitadas. Esto quiere decir que, las hip��tesis planteadas respecto a la predicci��n, la correcci��n basada en ��sta y la auto exposici��n, son factibles en aquellas situaciones identificadas a lo largo de la tesis pero que, sin embargo, no se puede garantizar que se cumplan de manera general. Por otra parte, se se��alan como trabajo de investigaci��n futuro algunas cuestiones nuevas y retos cient��ficos que aparecen a partir del trabajo presentado en esta tesis. ABSTRACT This thesis discusses the correction methods used to compensate the variation of lighting conditions in colour image and video applications. These variations are such that Computer Vision algorithms that use colour features to describe objects mostly fail. Three research questions are formulated that define the framework of the thesis. The first question addresses the similarities of the photometric behaviour between images of dissimilar adjacent surfaces. Based on the analysis of the image formation model in dynamic situations, this thesis proposes a model that predicts the colour variations of the region of an image from the variations of the surrounded regions. This proposed model is called the Quotient Relational Model of Regions. This model is valid when the light sources illuminate all of the surfaces included in the model; these surfaces are placed close each other, have similar orientations, and are primarily Lambertian. Under certain circumstances, a linear combination is established between the photometric responses of the regions. Previous work that proposed such a relational model was not found in the scientific literature. The second question examines whether those similarities could be used to correct the unknown photometric variations in an unknown region from the known adjacent regions. A method is proposed, called Linear Correction Mapping, which is capable of providing an affirmative answer under the circumstances previously characterised. A training stage is required to determine the parameters of the model. The method for single camera scenarios is extended to cover non-overlapping multi-camera architectures. To this extent, only several image samples of the same object acquired by all of the cameras are required. Furthermore, both the light variations and the changes in the camera exposure settings are covered by correction mapping. Every image correction method is unsuccessful when the image of the object to be corrected is overexposed or the signal-to-noise ratio is very low. Thus, the third question refers to the control of the acquisition process to obtain an optimal exposure in uncontrolled light conditions. A Camera Exposure Control method is proposed that is capable of holding a suitable exposure provided that the light variations can be collected within the dynamic range of the camera. Each one of the proposed methods was evaluated individually. The methodology of the experiments consisted of first selecting some scenarios that cover the representative situations for which the methods are theoretically valid. Linear Correction Mapping was validated using three object re-identification applications (vehicles, faces and persons) based on the object colour distributions. Camera Exposure Control was proved in an outdoor parking scenario. In addition, several performance indicators were defined to objectively compare the results with other relevant state of the art correction and auto-exposure methods. The results of the evaluation demonstrated that the proposed methods outperform the compared ones in the most situations. Based on the obtained results, the answers to the above-described research questions are affirmative in limited circumstances, that is, the hypothesis of the forecasting, the correction based on it, and the auto exposure are feasible in the situations identified in the thesis, although they cannot be guaranteed in general. Furthermore, the presented work raises new questions and scientific challenges, which are highlighted as future research work.

Reconfigurable video code (RVC) : aplicaci��n para automatizar la asignaci��n de procesos RUC a los n��cleos de un sistema multiprocesador

El esquema actual que existe en el ��mbito de la normalizaci��n y el dise��o de nuevos est��ndares de codificaci��n de v��deo se est�� convirtiendo en una tarea dif��cil de satisfacer la evoluci��n y dinamismo de la comunidad de codificaci��n de v��deo. El problema estaba centrado principalmente en poder explotar todas las caracter��sticas y similitudes entre los diferentes c��decs y est��ndares de codificaci��n. Esto ha obligado a tener que redise��ar algunas partes comunes a varios est��ndares de codificaci��n. Este problema origin�� la aparici��n de una nueva iniciativa de normalizaci��n dentro del comit�� ISO/IEC MPEG, llamado Reconfigurable Video Coding (RVC). Su principal idea era desarrollar un est��ndar de codificaci��n de v��deo que actualizase e incrementase progresivamente una biblioteca de los componentes, aportando flexibilidad y la capacidad de tener un c��digo reconfigurable mediante el uso de un nuevo lenguaje orientado a flujo de Actores/datos denominado CAL. Este lenguaje se usa para la especificaci��n de la biblioteca est��ndar y para la creaci��n de instancias del modelo del decodificador. M��s tarde, se desarroll�� un nuevo est��ndar de codificaci��n de v��deo denominado High Efficiency Video Coding (HEVC), que actualmente se encuentra en continuo proceso de actualizaci��n y desarrollo, que mejorase la eficiencia y compresi��n de la codificaci��n de v��deo. Obviamente se ha desarrollado una visi��n de HEVC empleando la metodolog��a de RVC. En este PFC, se emplean diferentes implementaciones de est��ndares empleando RVC. Por ejemplo mediante los decodificadores Mpeg 4 Part 2 SP y Mpeg 4 Part 10 CBP y PHP as�� como del nuevo est��ndar de codificaci��n HEVC, resaltando las caracter��sticas y utilidad de cada uno de ellos. En RVC los algoritmos se describen mediante una clase de actores que intercambian flujos de datos (tokens) para realizar diferentes acciones. El objetivo de este proyecto es desarrollar un programa que, partiendo de los decodificadores anteriormente mencionados, una serie de secuencia de v��deo en diferentes formatos de compresi��n y una distribuci��n est��ndar de los actores (para cada uno de los decodificadores), sea capaz de generar diferentes distribuciones de los actores del decodificador sobre uno o varios procesadores del sistema sobre el que se ejecuta, para conseguir la mayor eficiencia en la codificaci��n del v��deo. La finalidad del programa desarrollado en este proyecto es la de facilitar la realizaci��n de las distribuciones de los actores sobre los n��cleos del sistema, y obtener las mejores configuraciones posibles de una manera autom��tica y eficiente. ABSTRACT. The current scheme that exists in the field of standardization and the design of new video coding standards is becoming a difficult task to meet the evolving and dynamic community of video encoding. The problem was centered mainly in order to exploit all the features and similarities between different codecs and encoding standards. This has forced redesigning some parts common to several coding standards. This problem led to the emergence of a new initiative for standardization within the ISO / IEC MPEG committee, called Reconfigurable Video Coding (RVC). His main idea was to develop a video coding standard and gradually incrementase to update a library of components, providing flexibility and the ability to have a reconfigurable code using a new flow -oriented language Actors / data called CAL. This language is used for the specification of the standard library and to the instantiation model decoder. Later, a new video coding standard called High Efficiency Video Coding (HEVC), which currently is in continuous process of updating and development, which would improve the compression efficiency and video coding is developed. Obviously has developed a vision of using the methodology HEVC RVC. In this PFC, different implementations using RVC standard are used. For example, using decoders MPEG 4 Part 2 SP and MPEG 4 Part 10 CBP and PHP and the new coding standard HEVC, highlighting the features and usefulness of each. In RVC, the algorithms are described by a class of actors that exchange streams of data (tokens) to perform different actions. The objective of this project is to develop a program that, based on the aforementioned decoders, a series of video stream in different compression formats and a standard distribution of actors (for each of the decoders), is capable of generating different distributions decoder actors on one or more processors of the system on which it runs, to achieve greater efficiency in video coding. The purpose of the program developed in this project is to facilitate the realization of the distributions of the actors on the cores of the system, and get the best possible settings automatically and efficiently.

Development of a human movement monitoring system based on wearable devices

It is essential to remotely and continuously monitor the movements of individuals in many social areas, for example, taking care of aging people, physical therapy, athletic training etc. Many methods have been used, such as video record, motion analysis or sensor-based methods. Due to the limitations in remote communication, power consumption, portability and so on, most of them are not able to fulfill the requirements. The development of wearable technology and cloud computing provides a new efficient way to achieve this goal. This paper presents an intelligent human movement monitoring system based on a smartwatch, an Android smartphone and a distributed data management engine. This system includes advantages of wide adaptability, remote and long-term monitoring capacity, high portability and flexibility. The structure of the system and its principle are introduced. Four experiments are designed to prove the feasibility of the system. The results of the experiments demonstrate the system is able to detect different actions of individuals with adequate accuracy.

High-quality region-based foreground segmentation using a spatial grid of SVM classifiers

This paper presents a novel background modeling system that uses a spatial grid of Support Vector Machines classifiers for segmenting moving objects, which is a key step in many video-based consumer applications. The system is able to adapt to a large range of dynamic background situations since no parametric model or statistical distribution are assumed. This is achieved by using a different classifier per image region that learns the specific appearance of that scene region and its variations (illumination changes, dynamic backgrounds, etc.). The proposed system has been tested with a recent public database, outperforming other state-of-the-art algorithms.

Perceptual distortion modeling for side-by-side 3D video delivery

A frame-level distortion model based on perceptual features of the human visual system is proposed to improve the performance of unequal error protection strategies and provide better quality of experience to users in Side-by-Side 3D video delivery systems.

Fast 2D to 3D conversion using a clustering-based hierarchical search in a machine learning framework

Automatic 2D-to-3D conversion is an important application for filling the gap between the increasing number of 3D displays and the still scant 3D content. However, existing approaches have an excessive computational cost that complicates its practical application. In this paper, a fast automatic 2D-to-3D conversion technique is proposed, which uses a machine learning framework to infer the 3D structure of a query color image from a training database with color and depth images. Assuming that photometrically similar images have analogous 3D structures, a depth map is estimated by searching the most similar color images in the database, and fusing the corresponding depth maps. Large databases are desirable to achieve better results, but the computational cost also increases. A clustering-based hierarchical search using compact SURF descriptors to characterize images is proposed to drastically reduce search times. A significant computational time improvement has been obtained regarding other state-of-the-art approaches, maintaining the quality results.