Proyecto Fin De Carrera

In this M.Sc. Thesis, a biometric recognition system based on images simulated at the millimeter wave band of the electromagnetic spectrum is implemented and tested. In particular, body geometry information obtained from the silhouette of the body is used to determine the user's identity. In order to achieve this goal, firstly, some existing human body imaging technologies working beyond the visible spectrum are revised. Specifically the X-ray, infrared, millimeter and submillimeter wave imaging technologies are considered. The benefits and drawbacks of each technology as well as their biometric applications are also studied. Secondly, the generation of a new database is carried out. This database, called BIOGIGA, is composed of simulated images of people at 94GHz (within the millimeter wave band). The generation of this database was motivated by the lack of public databases of images of people at the mentioned spectral band. This lack is due to the recent development of imaging systems at that band and to the privacy concerns these images present. After that, the implementation of a biometric system based on the images of BIOGIGA is performed, including a feature selection module that considerably improves the error rates of the system. The main novelty of the proposed system with respect to another approaches already published, is that it is based on different geometric measures extracted from the silhouette of the body, such as the height, the wingspan, the waist width, the chest width, etc. This highly reduces the computational cost of the system. In the experimental section, by performing several experiments, the performance of the system under different conditions is evaluated. The effects of different conditions on the performance of the system are analyzed. Lastly, once the system is tested, the conclusions drawn throughout the work are presented together with the future work proposals. Gracias al resto de ATVSianos, en especial a: Dani (¿por qué vais vestidos?), Galbally (mujer de bellas piernas), Javi Manga (Javi, no somos unos alcohólicos!), Iñaki Perri (com-pañero de pádel y NYC host), Javi Pájaro (el más cuerdo del lab, no estantería!), Marta (con sus pócimas y Hello Kitty dominará el mundo, no dejes que Peter y RunVera te hinchen a pelotazos ni frutazos!), Ruifang (OMG! You know you are my hero, you canNOT dissapear now ;), Ram (I want you to speak Spanish. See you in all the ATVS-parties! :P), Mary Gates (crack!, lo que ha ganado USA, la ciencia y la …