The PLUS Multi-Sensor and Longitudinal Fingerprint Dataset: An Initial Quality and Performance Evaluation

In order to assess longitudinal effects in fingerprint biometrics several studies have been carried out the last decade. Almost all of these investigations focused on non-public forensic datasets because there is hardly any publicly available database allowing experiments and a corresponding analysis aspects often infeasible as contained samples are either captured one session only or with short time interval few weeks between sessions. With this work new dataset (108,106 samples) introduced which can be used for inter-session (longitudinal) well inter-sensor investigations. A total ten different capturing devices based distinct sensing technologies (optical, capacative, thermal, multispectral) utilized imprints fingers each 50 participating volunteers were acquired at four time-separated sessions over 2 years. This enables cross-device evaluations. Additionally, first quality recognition performance evaluation deploying state-of-the-art minutiae schemes NFIQ 2.0 measure presented.