Systematic evaluation methodology for fingerprint-image quality assessment techniques

Abstract

We propose a systematic methodology to assess fingerprintimage quality indices. Our approach is able to identify specific weaknesses or strengths of quality indices with respect to certain types of quality impairments. For that purpose, we suggest to employ a tool able to simulate a wide class of acquisition conditions and qualities, applicable to any given dataset and also of potential interest in forensic analysis. As an example for such a tool, StirMark image manipulations (as being developed in the context of watermarking robustness assessment) are applied to fingerprint data in this work, to generate corresponding test data, thereby interpreting certain image manipulations as being highly related to realistic fingerprint acquisition conditions. Experimental results document the usefulness of the approach and report significant inter-data set and inter-fingerprint matcher variability for the quality indices investigated.