Improving Dormant Matching Accuracy through Multiple Features

Abstract

Dormant Fingerprint Identification is of critical importance to Verification
agencies in identifying suspects. Dormant Fingerprints are inadvertent impressions left by
fingers on surfaces of object. While tremendous progress has been made in plain and rolled
fingerprint matching, Dormant fingerprint matching continues to be a difficult problem. We
study a number of features and combine multiple features at all three levels, including
ridge flow map, ridge quality map and ridge skeleton. Feature extraction and matching
algorithms are developed for each type of features. Relative contribution of each feature
towards the overall matching accuracy is evaluated by incrementally adding features to
baseline features (minutiae). Based on extensive experiments, all the extended features
lead to some improvement in Dormant matching accuracy.