Fingerprint skin moisture impact on biometric performance

Dry or wet fingers have an impact on biometric performance and the measurement of skin characteristics is important for actionable quality. We collected a ground truth fingerprint skin moisture dataset containing 6600 samples from 33 subjects using 5 optical sensors, two claimed to work well in extreme skin moisture conditions. Controlling conditions by applying cream, alcohol and water we performed objective measurements of skin moisture prior to each fingerprint capture. Analyzing the relationship between measured moisture and biometric performance across five sensors and comparison scores from 3 commercial providers confirms claims to resistance to excessive moisture. Results indicate that sensors resistant to excessive moisture perform worse on dry fingerprints than sensors not resistant to excessive moisture. The collected dataset is made available as an in-house dataset.