Elena Sanchez Nielsen

La deteccion y el seguimiento visual de objetos moviles en escenas dinamicas constituyen problemas centrales tratados en el area de investigacion de vision artificial. Estos problemas han sido abordados a lo largo de las dos ultimas decadas segun diferentes aproximaciones, pero se han ido encontrando con el cuello de botella resultante de asegurar por una parte, la ejecucion de los computos necesarios para llevar a cabo el proceso dentro de las limitaciones temporales que plantean las aplicaciones de vision en tiempo real y por otra parte, asegurar el robustecimiento del sistema resultante para llevar a cabo el proceso en entornos complejos no controlados y donde interactuan objetos naturales. Todo ello ha dado lugar a que aun hoy en dia siga siendo este un conjunto de problemas abierto en espera de soluciones robustas y rapidas. El objetivo de la presente tesis es el constituir una contribucion al problema del seguimiento precategorico de objetos deformables utilizando primitivas de contorno. Para ello, se parte de la consideracion de que los objetos a seguir son en general tridimensionales y pueden moverse con todos sus grados de libertad respecto al observador, que en general sera un sistema artificial con o sin egomovimiento. Los grados de libertad mencionados incluiran la deformacion del objeto, caso de que este no sea rigido como es el cuerpo humano. En base al modelo de presentacion visual establecido por Edelman, en este trabajo se analizan y ofrecen soluciones plausibles para cada uno de los siguientes aspectos fundamentales implicitos en el seguimiento visual precategorico: representacion del modelo, deteccion y comparacion, estructuras deformables, la problematica del tiempo real y una propuesta para el seguimiento de objetos deformables

Abstract: This paper introduces a fast and feasible method for the collection of hand gesture samples. Currently, there are not solid reference databases and standards for the evaluation and comparison of developed algorithms in hand posture recognition, and more generally in gesture recognition. These are two important issues that should be solved in order to improve research results. Unlike previous hand image datasets, which creation usually involves many different people, sceneries and light conditions, we propose a simplified method that requires just a single person ’ hand being recorded in a controlled light environment. Our method allows the generation of thousands of heterogeneous samples within hours, thus saving time and people’s efforts. The resulting dataset has been tested with a cascade classifier, although it may be used by most pattern recognition systems, and compared with a classical dataset obtaining similar results. 1

Una Arquitectura orientada a Servicios (SOA) es una aproximacion que esta directamente relacionada con el concepto de interoperabilidad entre sistemas de informacion y que permite disenar e implementar dichos sistemas, basados en componentes software denominados servicios. A su vez, SOA significa conceptos diferentes en funcion de la persona que haga uso de dicho termino: (1) analista de negocio, (2) arquitecto de software y (3) programador. En este articulo, se describen los objetivos, contenidos teoricos y practicos, asi como la metodologia docente de una asignatura de Arquitecturas orientadas a Servicios en el contexto del EEES que unifica las diferentes visiones sobre SOA en una vision holistica, haciendo especial enfasis tanto en el sector de la industria y administracion publica, como en el sector de la investigacion. Dicha asignatura se ha impartido como asignatura obligatoria durante los cursos academicos 2008/09 y 2009/10 en el Programa Oficial de Posgrado en Informatica, c...

Los objetivos, contenidos teoricos y practicos relacionados con las asignaturas de Inteligencia Artificial y Agentes en la Ingenieria Superior de Informatica, suelen ser comunes en los diferentes planes docentes de las diversas titulaciones que conforman las diferentes universidades espanolas, utilizandose tradicionalmente, herramientas provenientes del mundo academico y de la investigacion, para llevar a cabo el desarrollo e implementacion de los contenidos practicos de dichas asignaturas. Esta situacion puede conducir a que en muchas ocasiones, los contenidos practicos sean excesivamente academicos y artificiales. En este articulo, se describe como se puede introducir la tecnologia de desarrollo de Google en los contenidos practicos de las asignaturas de Inteligencia Artificial y Agentes Inteligentes para: (1) acercar los contenidos academicos del estudiante hacia herramientas reales y necesidades especificas demandadas por la industria, (2) concebir los resultados del desarrollo ...

En este articulo, se describe una metodologia docente basada en la elaboracion y produccion de contenidos audiovisuales, por parte de los estudiantes, como herramienta para mostrar los resultados de las actividades desarrolladas. La metodologia planteada pretende fomentar el aprendizaje centrado en una actitud activa del estudiante, asi como la participacion e interes en los resultados elaborados por los otros estudiantes. Se muestra como la elaboracion de contenidos audiovisuales producidos por los estudiantes, y su evaluacion entre los mismos promueve que el estudiante encuentre motivacion en su propio proceso de aprendizaje. La experiencia docente y los resultados obtenidos tanto en una asignatura de grado, como en una asignatura de master de la ingenieria informatica muestran como se fomenta la participacion e interaccion, trabajo cooperativo, el analisis critico, y la creatividad. Asimismo, el 98% y 100% de los estudiantes de grado y master finalizaron respectivamente sus activ...

Tracking moving objects based on a Hausdorff approach can be formulated in terms of a matching process between two sets of edge points extracted from the object model to be localized and the corresponding frame of the image sequence. However, no information about the global measures of the object feature positions has been considered to carry out the matching process

Many vision problems require computing fast template motion in dynamic scenes. These problems can be formulated as exploration problems and thus can be expressed as a search into a state space based representation approach. However, these problems are hard to solve because they involve search through a high dimensional space. In this paper, we propose a heuristic algorithm through the space of transformations for computing target 2D motion. Three features are combined in order to compute efficient motion: (1) a quality of function match based on a holistic similarity measurement, (2) Kullback-Leibler measure as heuristic to guide the search process and (3) incorporation of target dynamics into the search process for computing the most promising search alternatives. The paper includes experimental evaluations that illustrate the efficiency and suitability for real-time vision based tasks.

ABSTRACT An architecture model of inter-administrative cooperation for Spanish autonomous parliaments based on different layers of functionality is presented in this paper through the WebSpanishParliament project (WebSP) in order to perform one of the major initiatives of COPREPA (conference ...

... 24 E. Sánchez-Nielsen, A. Padrón-Ferrer, and F. Marreo-Estévez ... to date are: (1) FIPA specifications [11], a set of specifications intended to sup-port the interoperability between heterogeneous agent-based systems; and (2) an agent development framework, called JADE [ ...

This paper introduces a fast and feasible method for the collection of hand gesture samples. Currently, there are not solid reference databases and standards for the evaluation and comparison of developed algorithms in hand posture recognition, and more generally in gesture recognition. These are two important issues that should be solved in order to improve research results. Unlike previous hand

The Distance Transform is a powerful tool that has been used in many computer vision tasks. In this paper, the use of relevant maxima in distance transform’s medial axis is proposed as a method for fast image data reduction. These disc-shaped maxima include morphological information from the object they belong to, and because maxima are located inside homogeneous regions, they

... Luis Antón-Canalıs1, Mario Hernández-Tejera1, and Elena Sánchez-Nielsen2 ... Tomasi and Manduchi's Bilateral filter [9] and Perona and Malik's anisotropic diffusion [8] achieve this through intensity/color and gradient dependent computations between pixels. ...

A solution based on a Swarm Intelligence metaphor with a prey-predator scheme is proposed for real time ob-ject tracking in video sequences, which is a basic process in multiple computer vision tasks. Swarm predator par-ticles fly on a Boid-like fashion over image prey pixels, ...

Abstract. This paper presents a system for hand posture recognition that works with colour video streams under varying light conditions for human-machine interaction in vision-based interface tasks. No initializa-tion of the system is required and no user dependence is involved. ...

ABSTRACT The use of Google technology in artificial intelligence courses gives new opportunities to focus the academic topics and learning objectives in order to meet the needs of students and computer science curricula. With the integration of Google developer toolkits, we can transform the class in order to use industry software tools as a motivating topic and as the domain for real-world software projects. This paper describes the methodology and motivating results using Google developer toolkits in teaching and practicing computer applications, which are based on artificial intelligence theory and techniques.

Delivery of personalized multimedia content is an important challenge for versatile video podcast environments. In this paper, we focus on the opportunities given by MPEG-7 metadata standard for content delivery in terms of user preferences when podcasting technology is used. By means of MPEG-7, multimedia content can be described at several levels of granularity and abstraction. We propose the use of this description in order to produce and publish personalized information according to user's demand on podcasts environments and to compute at a dynamic way the different feeds of the podcast information system when it is demanded by users. We describe this approach for the motivating scenario of audiovisual content for legislative assemblies.

Current and geographically distributed legislative institutions demand exchange of information between related entities. However, existing legislative applications are composed of autonomous and heterogeneous components. In this paper, we describe an e-services model for Spanish legislative distributed assemblies. Our approach addresses the following issues: (I) a scalar and interoperable framework that allows incorporating new institutions and e-government services in a dynamic