A Study of Friction Ridge Distortion Effect on Automated Fingerprint Identification System - Database Evaluation

摘要

Fingerprint identification is an important part of forensic science (e.g. criminal investigations or identity verification). Friction ridge impressions left at the crime scene can be affected by the nonlinear distortion due to elasticity of the skin, pressure changes or finger movement during deposition. These deformations affect relative distances between fingerprint features such as minutiae point, ridge frequency and orientation, which eventually leads to difficulties in establishing a positive match between impressions of the same finger. In this study we present preliminary results of the impact of fingerprint friction ridge distortion on NBIS Bozorth3 fingerprint matching algorithm. For this purpose special fingerprint database was developed. The database contained 5175 prints obtained from 40 volunteers. Experimental results reveal that the some types of fingerprint distortion (especially movement to right and left) impacts the recognition performance. The results of our studies can be used in future work on statistical friction ridge analysis and fingerprint algorithms robust to distortions.