Study On Close Non-match Of Loops In Millions Of Database

The uniqueness and persistence is not only the characteristic of fingerprints,but also the basic of fingerprint identification.Because of the characteristic,fingerprint evidence has been widely used in the criminal investigation.With few identification error cases exposed,people began to realize that two different fingerprint patterns maybe very similar in some area of fingerprints.At the same time,the automatic fingerprint identification system(AFIS)database size increasing is also a challenge for fingerprint identification.The difference of experience and cognitive ability between examiners become more obvious when fingerprints from same source appear and close non-matching fingerprints generate after AFIS searching.The study can raise the awareness of the risks encountering close non-matching fingerprints and provide the data for examiners.This study looked for close non-matching fingerprints through selecting 8?10?12 details in the delta,center and branch area of 80 loops in AFIS database.Then it analyzed if the number and the area of the details influence the occurrence frequency and position of fingerprints from same source.Meanwhile the factors influenced close non-matching fingerprints were also studied.The results showed that the number and the area of the details has an effect on fingerprint come from the same source.The number and the area of the details,finger position influence non-matching fingerprints generating.The fingerprint come from left or right hand did not affect close non-matching fingerprints and left loop or right loop did not neither.Finally,the study evaluated scientific of fingerprint evidence by probability of coincidence and likelihood ratio,which provide scientific supports for fingerprint identification basis theoretical.The results of this study can provide some suggestions for AFIS query,help the developers of the AFIS system to improve the query methods and develop engineering controls within AFIS.These risk assessments may be used to help examiners better understand the minutiae and improve fingerprint recognition ability.This paper also provides some reference for the research on basic theory of fingerprint identification and makes a foundation for close non-matching fingerprints through the probabilistic method,which will promote fingerprint evidence evaluation from experience to science in forensic science.