作者: Wei Zhou , Jiankun Hu , Song Wang , Ian Petersen , Mohammed Bennamoun
DOI: 10.1007/978-3-319-11698-3_29
关键词:
摘要: Fingerprint indexing is a process of pre-filtering the template database before matching. The most common features used for fingerprint are based on minutiae triplets. In this paper, we investigated performance some commonly triplets and proposed to combine these with novel indexing. Experiments FVC 2000 DB2a 2002 DB1a show that method can perform better than state-of-the-art schemes full indexing, meanwhile, experimental results NIST SD 14 improved significantly after new added feature space, fairly good even partial