Classificação de impressões digitais

Carregando...
Imagem de Miniatura

Data

2011

Autores

Mizobuchi Júnior, Alvaro Toshiyuki [UNESP]

Título da Revista

ISSN da Revista

Título de Volume

Editor

Universidade Estadual Paulista (Unesp)

Resumo

The biometric characteristics have been used increasingly as a way to identify an individual, mainly for security reasons. Among them, the fingerprint is the most used biometric characteristic around the world, because it is relatively simple and very efficient. In this scene, there was a significant increase in the size of databases containing information on fingerprints, necessary to perform the recognition of a person. The task of classifying them beforehand has become extremely important as it reduces dramatically the size of the problem during a search, because it is not necessary to go through the whole database. Considering its importance, in the last thirty years, many techniques have been developed to try to increase the efficiency of the classification process. This project followed the rules-based approach and the Software Development Kit (SDK) VeriFinger 6.1 was used to assist in the detection of cores and deltas. Additionally, the classification was also implemented by means of directional map and the Poincar´e index. To make the experiments, the number four database from the National Institute of Standards and Technology (NIST) was used, which is a standard in this area
As características biométricas têm sido utilizadas cada vez mais como forma de identificação de um indivíduo, principalmente por motivos de segurança. Dentre elas, a impress ão digitalé a característica biométrica mais utilizada em todo o mundo por ser relativamente simples e muito eficiente. Diante deste cenário, houve um aumento expressivo no tamanho das bases de dados contendo informações sobre as impressões digitais, necessárias para realizar o reconhecimento de um sujeito. Assim, a tarefa de classificá-las previamente tornou-se extremamente importante, já que reduz drasticamente a dimensão do problema durante uma busca, não sendo necessário percorrer toda essa base de dados. Considerando a sua importância, nos últimos trinta anos, muitas técnicas foram desenvolvidas para se tentar aumentar a eficiência do processo de classificação. Neste projeto seguiu-se a abordagem baseada em regras e foi utilizado o Software Development Kit (SDK) Verifinger 6.1 para auxiliar na detecção de núcleos e deltas. Adicionalmente, também foi implementado a classificação por meio do mapa direcional e do cálculo do Indice de Poincaré. Para a realização dos experimentos, foi feito uso da base de dados número quatro do National Institute of Standards and Technology (NIST), que ´e um padrão da área

Descrição

Palavras-chave

Computação, Biometria, Reconhecimento de padrões, Processamento de imagens, Datiloscopia

Como citar

MIZOBUCHI JÚNIOR, Alvaro Toshiyuki. Classificação de impressões digitais. 2011. 1 CD-ROM. Trabalho de conclusão de curso (bacharelado - Ciência da Computação) - Universidade Estadual Paulista, Faculdade de Ciências e Tecnologia, 2011.