Extracción de características identificativas en plantillas de minucias mediante la estructura compleja

• Main Authors: Ramón Santana Fernández, Adrian A. Machado Cento, Vivian Estrada Senti
• Format: Article
• Language: Spanish
• Published: Universidad de Ciencias Informáticas 2015-10-01
• Series: Revista Cubana de Ciencias Informáticas
• Subjects: minutiae structure; information representation; minutiae template; template protection
• Online Access: http://rcci.uci.cu/index.php?journal=rcci&page=article&op=view&path[]=1167&path[]=383

Identifying people using biometric features such as ngerprints, it is considered a secure authentication me- chanism. Cryptographic systems proposed for the protection template ngerprint minutiae need to perform the alignment process of the biometric templates due to variability of data in dierent samples of the same feature. Alignment models proposed in the literature base their operation on the selection of a focal point, the formation of structures of minutiae or detection of the singularities of the ngerprint. These features may or may not be present in the ngerprint, they may vary from one sample to another and may be selected wrongly, conse- quently resulting in loss of performance of authentication system. This research aims at developing a method of representing the information contained in the minutiae templates and identifying characteristics extraction, rotation and translation invariant, which enables the analysis of the characteristics of the minutiae obtained locally and overall, to lessen the impact of the elimination of real minutiae and the inclusion of false minutiae, using the combination of two structures minutiae. The process consists of 3 stages: formation of the complex structure of minutiae extraction invariant to rotation and translation characteristics from triplets and classi- cation of the extracted features. The comparison process is performed by calculating similarity between each complex structure.