Publication: Mecanismos de captura y procesado de imágenes de venas para identificación personal
Loading...
Identifiers
Publication date
2011-07
Defense date
2011-11-07
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
Actualmente la identificación biométrica para el acceso a recintos o servicios está cobrando bastante importancia. La biometría es una tecnología que, además de realizar funciones de reconocimiento, aporta mayores niveles de seguridad que otros métodos de identificación externos, al llevar intrínsecamente la información en el cuerpo. La identificación biométrica vascular ofrece un futuro muy prometedor en ambos aspectos. La mayoría de los sistemas biométricos se basan en sistemas hardware (HW) para la captura de muestras del usuario más una solución software (SW) para el procesado de las mismas. La tarea de la investigación en biometría vascular hoy en día es crear sistemas fiables y seguros para obtener soluciones aplicables. En esta Tesis, hemos propuesto dos importantes contribuciones a la identificación mediante esta modalidad biométrica: 1. Un sistema hardware novedoso y económico para la captura de patrones y muestras del usuario, tanto de palma como de muñeca, que ofrece imágenes de una calidad suficiente para su procesado posterior. Además, a diferencia de los trabajos previos realizados, las especificaciones y diseño de este sistema se ofrecen abiertamente a la comunidad científica. 2. Un sistema de procesamiento de las imágenes de las venas extraídas, basado en las conocidas función Gaussiana y función de Convolución, y un nuevo sistema de toma de decisiones, basado en una comparación orientable de las líneas extraídas apoyado con el dato obtenido del ancho de la muñeca de cada usuario. Esto también se ofrece a la comunidad científica, mostrando todas las fases del proceso, al contrario de lo que ocurre con las escasas publicaciones existentes que se analizarán a lo largo del presente documento. Para demostrar la viabilidad de esta propuesta se ha desarrollado e implementado un sistema HW y SW completo, y se ha probado finalmente con 100 usuarios de diferentes edades, razas y sexo, obteniendo los resultados mostrados al final de esta Tesis que confirman la viabilidad futura de la solución propuesta. Finalmente se indican las líneas de investigación futuras para mejorar la solución y obtener mejores resultados, así como para obtener una mayor comodidad a la hora de la captura de muestras del usuario. ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Biometric identification for the compartment or service access is more and more important nowadays. Biometrics is a technology which brings higher security leves than other external identification methods because the information is inherently inside the human body, apart from providing recognition functions. Vascular biometric identification offers a very hopeful future in both subjects. Most of the biometric systems are based on hardware systems (HW) in order to take user samples in addition to software systems (SW) so as to process these ones. The investigation task in vascular biometrics is focused on creating reliable and secure systems at this moment to obtain relevant solutions. We propose in this PhD thesis two important contributions for the identification using the above mentioned biometric modality: 1. A novel and economic hardware system in order to capture user patterns and samples, palm and wrist both, which offers images with enough quality for its further processing. Furthermore, unlike the previous found works, the design and specifications of this system are available openly to the scientific community. 2. An image processing system for the extracted veins images, based on the well-known Gauss and Convolution functions, and a novel decision making system, based on an orientation comparison of the extracted lines with the support of the wrist width value obtained for every user. All of this is also available to the scientific community, explaining every step in the whole process, unlike what happens with the limited existing scientific publications, which will be discussed along the present document. A complete HW and SW system has been designed and developed so as to demonstrate the viability of these works, and this one has been finally tested with 100 users from different ages, races and sex, obtaining the results shown at the end of this PhD thesis, which confirm the future viability of the proposed solution. Finally, future investigation points are given to enhance the solution and to get better results, as well as to obtain a major comfort for the user sample captures.
Biometric identification for the compartment or service access is more and more important nowadays. Biometrics is a technology which brings higher security leves than other external identification methods because the information is inherently inside the human body, apart from providing recognition functions. Vascular biometric identification offers a very hopeful future in both subjects. Most of the biometric systems are based on hardware systems (HW) in order to take user samples in addition to software systems (SW) so as to process these ones. The investigation task in vascular biometrics is focused on creating reliable and secure systems at this moment to obtain relevant solutions. We propose in this PhD thesis two important contributions for the identification using the above mentioned biometric modality: 1. A novel and economic hardware system in order to capture user patterns and samples, palm and wrist both, which offers images with enough quality for its further processing. Furthermore, unlike the previous found works, the design and specifications of this system are available openly to the scientific community. 2. An image processing system for the extracted veins images, based on the well-known Gauss and Convolution functions, and a novel decision making system, based on an orientation comparison of the extracted lines with the support of the wrist width value obtained for every user. All of this is also available to the scientific community, explaining every step in the whole process, unlike what happens with the limited existing scientific publications, which will be discussed along the present document. A complete HW and SW system has been designed and developed so as to demonstrate the viability of these works, and this one has been finally tested with 100 users from different ages, races and sex, obtaining the results shown at the end of this PhD thesis, which confirm the future viability of the proposed solution. Finally, future investigation points are given to enhance the solution and to get better results, as well as to obtain a major comfort for the user sample captures.
Description
Keywords
Biometría, Identificación biométrica vascular, Proceso de imágenes