Identificación y clasificación biométrica a gran escala basada en huellas dactilares y GPU
Metadatos
Afficher la notice complèteAuteur
Gutiérrez Pérez, Pablo DavidEditorial
Universidad de Granada
Departamento
Universidad de Granada. Departamento de Ciencias de la Computación e Inteligencia ArtificialMateria
Biometría Identificación Dactiloscopia GPU (Graphics Processing Units) Ordenación (Estadística) Tecnologías de la información y comunicación (TICs) Procesado de imágenes Reconocimiento óptico de formas (Informática)
Materia UDC
004 004.65 3304
Date
2017Fecha lectura
2017-06-01Referencia bibliográfica
Gutiérrez Pérez, P. D. Identificación y clasificación biométrica a gran escala basada en huellas dactilares y GPU. Granada: Universidad de Granada, 2017. [http://hdl.handle.net/10481/48506]
Patrocinador
Tesis Univ. Granada. Programa Oficial de Doctorado en Tecnologías de la Información y la Comunicación; Esta tesis doctoral ha sido desarrollada con la financiación del Programa de Becas de Formación de Personal Investigador del Ministerio de Economía y Comptetitividad, en su Resolución del 28 de Noviembre de 2012, bajo la referencia BES-2012-060450.Résumé
In this thesis we are going to study how GPU devices can be used to improve the performance of fingerprint matching methods and imbalanced classification strategies that are suitable to be introduced in a large scale AFIS. In order to do that, we first studied the characteristics of GPU devices and matching methods. Then, we proposed GPU-based designs for two of the most representative matching algorithms. After that, we studied how to solve the imbalanced learning problem in large databases by providing a scalable solution for one of the most well-known techniques to deal with imbalanced classification and the underlying classification algorithm that this technique uses.
Section 2 describes in detail the background of the main topics covered by the thesis: GPU devices and their characteristics (Section 2.1), fingerprint matching (Section 2.2) and classification and imbalanced data (Section 2.3).
Section 3 presents the justification of this doctoral dissertation, describing the problems addressed throughout the thesis. Section 4 marks the objectives pursued and Section 5 introduces the methodology used to achieve them. Section 6 summarizes the works that comprise this thesis, while Section 7 analyses their results in relation to the objectives. Section 8 presents the conclusions obtained during the work collected in this thesis and Section 9 points out future lines of work derived from the results obtained.