RT Generic
T1 Plataforma de Captura de Base de Datos Biométricos de acuerdo a LOPD
A1 Ortega Fernández, Inés
AB En los últimos años, la computación se ha introducido en ámbitos de la sociedad antesinimaginables. Hoy en día todos llevamos en nuestro bolsillo un dispositivo cientos de vecesmás potente que cualquier computadora de los años 80. No sólo eso, sino que podemosencontrar computadoras en nuestra muñeca, el coche o la nevera. El aumento del númerode dispositivos móviles ha provocado el surgimiento de nuevas formas de identificación yautenticación.Como parte de esta expansión, se han llevado tecnologías más propias de gobiernos einstituciones de alta tecnología a la vida cotidiana de los usuarios. Un claro ejemplo son lossistemas de reconocimiento biométrico: desbloqueamos nuestro móvil con la huella dactilar,el iris o mediante reconocimiento facial, accedemos a zonas de acceso restringido omostramos nuestra identidad en un aeropuerto.Estos avances han sido posibles en gran medida gracias al abandono del métodotradicional de adquisición y comparación de huellas dactilares en papel para dejar paso aluso de dispositivos electrónicos, los sensores de huella dactilar.El presente documento tiene como objetivo la documentación de una plataforma de capturade muestras biométricas, concretamente huellas dactilares, capaz de recoger muestras através de cinco sensores diferentes, analizando la calidad y procesando las muestrasobtenidas. Este tipo de plataformas automatizan enormemente el proceso de recogida demuestras biométricas, y por ende la evaluación de sistemas de reconocimiento biométrico.La plataforma permite tanto realizar el registro del usuario en el sistema (almacenando susdatos personales) como capturar las muestras biométricas necesarias. La captura demuestras puede realizarse de dos modos: para reclutamiento, o para adquisición. Lasmuestras obtenidas durante el reclutamiento actuan como referencia del usuario en laplataforma, mientras que las muestras obtenidas durante la fase de adquisición seráncomparadas con las referencias, determinando si pertenecen o no al mismo usuario.Este proyecto constituye el primer paso de un proyecto mayor, que busca realizar unaevaluación de rendimiento de cinco sensores de captura de huella dactilar llevada a cabopor el Grupo Universitario de Tecnologías de Identificación (GUTI) de la Universidad CarlosIII de Madrid. Las muestras y estadísticas de uso recogidas con esta plataforma, seránposteriormente procesadas y comparadas, lo que nos permitirá conocer el rendimiento decada uno de los sensores.
AB In the last few years, computing has been introduced in environments of the societyinconceivable before. Nowadays we all carry on our pockets devices much more powerfulthan the computers from the 80s. We can even find computers in our wrist, the car or thefridge. The increasing number of mobile devices has caused the raising of new ways ofidentification and authentication.As part of this expansion, new technologies have been taken to peoples’ daily life. Forexample, biometric recognition systems are now used to unlock our smartphone, using forexample fingerprint, facial or iris recognition.Biometrics is a field that studies in a statistically way different biology parameters. Biometricrecognition, more concretely, groups all those technologies that allow the identificationand/or verification of the identity of a subject from its morphological or behaviourcharacteristics.To be considered to be used for biometric recognition, a biological characteristic has toachieve some conditions [1]:Universality: can be found in any person.Uniqueness: has to be different enough between two persons.Continuance: may stay invariable along the time.Measurable: may be measured in a quantitative way.A fingerprint is the representation of the morphological surface of the epidermis of a finger.It has a serie of lines that, globally, are distributed in a parallel way. However these linesintersects and sometimes ends roughly. These points are technically called characteristicpoints of a fingerprint. To conclude if two fingerprints belong to the same person, thecharacteristic points of both fingerprints are compared.Fingerprints can be used in a biometric recognition system because achieves therequirements stated previously:Are universal, as every person has fingerprints.Are unique: every fingerprint is different in any person; its generation is not a geneticprocess, but random.Are perennial: they are formed during the sixth month of pregnancy, and stayinvariable during a person lifetime, growing proportionally to the finger size.Its acquisition can be done in a easy and fast way. Multiple different fingerprints canbe obtained from the same person, because each fingerprint from each finger isdifferent even for the same individual.In the field of fingerprints’ recognition systems, the improvement of the technology has beenmade by the abandonment of traditional methods of acquisition and comparison of samplesin benefit of the use of fingerprint sensors.Biometric recognition is a booming technology in the field of access control. Its main benefitin relation to traditional access control systems (like passwords) is that the user doesn’t haveto remember anything, or even carry an identity card. However, the implementation ofbiometric recognition systems can be more expensive, and requires a high technologicalcomponent.Generally, a biometric system can be divided into two stages: the enrolment and the use ofthe samples obtained:Enrolment:A serie of samples are taken from the user, and are processed to extract a pattern.This pattern will be stored, and will be used as the reference of the user in the system.One of the characteristics of this process is that usually is supervised by an operator,who controls that the data capture is done in a proper way.In addition, usually a minium quality threshold is established. This threshold can bebased on the operator criteria, or an algorithm can be used to quantify the quality ofthe sample.Use of the samples:Once the system has stored the user’s pattern, the user can use the system. Thesamples obtained during this stage will be compared against the pattern obtained inthe enrolment process. This comparison will result in a similarity score between thepatterns.If the comparison is done against one reference, to check if both samples belong tothe same individual, the process is called verification. If a sample is taken and asearch is performed along all the database to find to which user it belongs, theprocess is called identification.If the comparison success, the user can use the system (for example, access to arestricted area, or unlock its smartphone). However, if the comparison doesn’tsucceed, it means that an error was produced during the identification process, andthe task for what the identification was required can’t not be performed.The following figure shows the schema of fingerprint recognition systems.
YR 2017
FD 2017-07-06
LK http://hdl.handle.net/10016/27410
UL http://hdl.handle.net/10016/27410
LA spa
DS e-Archivo
RD 28 abr. 2024