Detecção de impressões digitais falsas no reconhecimento biométrico de pessoas

• Ano de defesa: 2015
• Autor(a) principal: Silva, Murilo Varges da [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Computação; Sistemas de detecção de intrusão (Medidas de segurança); Biometria; Datiloscopia; Computer science
• Link de acesso: http://hdl.handle.net/11449/136775; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/24-03-2016/000859947.pdf

In recent years, many biometrics traits have been proposed to biometric identification of people, among which stand out the face, iris, retina, hand geometry. However, fingerprint is still the most used feature in both commercial and government applications. The main applications we can mention are the identification of voters through electronic voting machines, border control and immigration through passports, access to banking services through biometric Automated Teller Machine (ATM), among others. With the increased use of these systems, attempts to attack also increase. Among the types of attacks that a biometric system based on fingerprint may suffer, presenting a false finger to the sensor is the most used technique by malicious people. This work aims to propose a robust method to detect fake fingerprints by incorporating information from the sweat pores. The proposed method was assessed on own database named UNESP Fingerprint Spoof Database (UNESP-FSDB), the analysis considered some factors that may influence method performance as: (i) fingerprint image resolution ; (ii) use of third-level characteristics (pores); (iii) finger pressure on the sensor surface; (iv) fingerprint image acquisition time; (v) finger moisture present in the moment of capture fingerprint. The results of the experiments showed that: (i) incorporating information of pores can increase the accuracy up to 9%; (ii) using increased pressure at the time of capture improves the performance; (iii) the moisture present in the finger can influence the image quality and accuracy of the proposed method