A new similarity digest search strategy applied to Minutia Cylinder-Codes for fingerprint identification
ResumoOne challenging problem on the fingerprint realm is the identification of individuals over large databases, where the most similar template must be found. Approximate matching is used in digital forensic investigations to deal efficiently with large amount of data. We think it can also be used to identify similar fingerprints with compact representations and become a promising technique to speed up searches. In this paper, we explore this hypothesis and present MCC-HBFT, a new fingerprint identification strategy based on the approximate matching technique HBFT and the state-of-the-art fingerprint representation model MCC. We show how MCC-HBFT identify fingerprints and outperforms a commonly used indexing strategy in some public databases of the field.
MOIA, Vitor Hugo Galhardo; HENRIQUES, Marco Aurélio Amaral. A new similarity digest search strategy applied to Minutia Cylinder-Codes for fingerprint identification. In: SIMPÓSIO BRASILEIRO EM SEGURANÇA DA INFORMAÇÃO E DE SISTEMAS COMPUTACIONAIS (SBSEG) , 2018 Anais do XVIII Simpósio Brasileiro em Segurança da Informação e de Sistemas Computacionais. Porto Alegre: Sociedade Brasileira de Computação, oct. 2018 . p. 99 - 112.