Rotationally Invariant Statistics for Examining the Evidence from the Pores in Fingerprints

Recent methodological advances in the processing of DNA evidence have begun to force a closer examination of assertions about the strength of other sorts of evidence. One traditional source of evidence is the fingerprint. Currently a print taken from a suspect is compared against a mark from a crime scene and a match declared using the judgement of an expert based on matching minutiae and the ridge patterns around these. However, such methods have proved difficult to quantify effectively. This has provoked the investigation of even finer features in the print and the mark. One set of such features are the many pores, located along the ridges of the fingerprint. Is it possible to supplement expert judgements associated with a match with a more automatic and quantitative measure of the strength of evidence, based on pore information? The results of this preliminary analysis suggest we can. Our methodology is relatively transparent, using common statistics for two sample comparisons of point patterns. The results discussed here concern the matching of inked prints using grey-level imaging and complement previous studies which tend to focus on the comparison of binarised images