Detecção de impressões digitais fraudulentas utilizando padrões mapeados localmente em multiescala

• Ano de defesa: 2015
• Autor(a) principal: Contreras, Rodrigo Colnago [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Computação - Matematica; Modelos matematicos; Sistemas de detecção de intrusão (Medidas de segurança); Biometria; Reconhecimento de padrões; Aprendizado do computador; Mathematical models
• Link de acesso: http://hdl.handle.net/11449/136662; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/24-03-2016/000859900.pdf

Applications which use fingerprint recognition are present in many contexts of everyday life, for example, transactions at ATMs, electronic registration point in companies and, since 2008, in the Brazilian electoral process. The fingerprint matching is usually made based on the relative positions of small patterns known as minutiae. Recent scientific researches show the possibility of making synthetic fingerprints, which have a copy of these minutiae in an authentic finger, from simple materials such as play dough, silicone, wood glue, among others. Faced with the imminent danger of fraud in biometric systems, a new line of research emerged: the detection of fake fingerprints (spoofs). Currently, they are in the literature some methods to minimize this problem, but this is still an open problem. The present study focuses on the detection of fake fingerprints using image processing and pattern recognition techniques. The main contribution of this study is a new technique with which the fakes one are detected from the analysis of spatial micro patterns extracted from the fingerprints. Moreover, it is developed in this work, the multi-scale version of the space micro pattern known as Local Mapped Pattern (LMP). The experiments carried out have shown that through the proposed technique, it is possible to achieve a higher detection performances with the results found state of the art in the specialized literature