Fingerprint Core and Delta Detection by Candidate Analysis

This paper presents a new reliable detection of core and delta in fingerprints by candidate analysis. The line search based method, which is one of the reliable approaches to detect core and delta, is improved aiming to realize the high positional precision and short computational time. In order to avoid the detection error, the three kinds of candidates, where the probability of core and delta existence becomes higher, are introduced. In the candidate analysis, the density of candidates is evaluated over the fingerprint. Also the structure of the extended relational graph generated from the directional image of the fingerprint is analyzed to find the number of core loop and delta loop. As same as the number of the core and delta loops, the connected candidate groups are picked up in the order of the candidate density. The center of each candidate group becomes core or delta. Based on several experimental results, the success rate of the core and delta detection can be achieved about ...

• Type: Conference
• Year: 2007
• Where: MVA
• Authors: Tomohiko Ohtsuka, Daisuke Watanabe, Hiroyuki Aoki
• Keywords: Candidate Analysis, Computer Vision, Delta Loops, MVA 2007, Search Based Method