|
|||||||
| |||||||
Evaluation Results - Extension and further Evaluation of the PLUS Multi-Sensor and Longitudinal Fingerprint Dataset | |||||||
This is "The Multimedia Signal Processing and Security Lab", short WaveLab, website. We are a research group at the Artificial Intelligence and Human Interfaces (AIHI)
Department of the University of Salzburg led by Andreas Uhl.
Our research is focused on Visual Data Processing and associated security questions. Most of our work is currently concentrated on Biometrics, Media Forensics and
Media Security, Medical Image and Video Analysis, and application oriented fundamental research in digital humanities, individualised aquaculture and sustainable wood
industry.
| |||||||
Evaluation Results for Extension and further Evaluation of the PLUS Multi-Sensor and Longitudinal Fingerprint DatasetIn the paper "Extension and further Evaluation of the PLUS Multi-Sensor and Longitudinal Fingerprint Dataset" we present an extension to the previously published PLUS-MSL-FP dataset (which can be found here: PLUS-MSL-FP Database - Original Version), which is a comprehensive fingerprint dataset, providing a publicly available baseline for further investigations on the aspects of FP aging. In the above mentioned paper we extended the dataset by an additional, fifth session which has been captured 3 years after the last session of the initial dataset. This extended version of the dataset was evaluated in the same way as the original one. To be in-line with the principles of reproducible research the detailed results of this evaluation regarding fingerprint quality as well as fingerprint recognition performance can be found here. AbstractInter-session or template aging related effects in fingerprint biometrics have been discussed controversially in the last decade. The lack of publicly available fingerprint databases covering longer inter-session time intervals (months to years) is the major problem limiting further research. In a previous work the PLUS MSL FP dataset was introduced, containing 108,106 fingerprint samples collected from 50 subjects in 4 sessions over a time span of two years, involving 10 different different capturing devices (optical, capacitive, thermal and multispectral ones). This work extends the PLUS MSL FP dataset by an additional, 5th session, captured 3 years after the 4th one. The additional session is evaluated regarding its fingerprint image quality (NFIQ 2.2) as well as its recognition performance (employing minutiae based fingerprint recognition schemes). The results confirm the previous findings: a trend towards lower inter-session performance compared to the intra-session one, most likely due to changes in subjects' behavior across different sessions. An additional evaluation on a small short-term inter-session fingerprint dataset revealed that this trend is not there for subjects used to handling fingerprint capturing devices. Reference[Kauba24a ] Extension and further Evaluation of the PLUS Multi-Sensor and Longitudinal Fingerprint Dataset In Proceedings of the 12th International Workshop on Biometrics and Forensics (IWBF'24), pp. 1-6, Enschede, Netherlands, April 11 - April 12, 2024
Data SetPLUS-MSL-FP-v2 DatabaseThe second (extended) version of the PLUS Multi-Session and Longitudinal Fingerprint Database (PLUS-MSL-FP-v2) is a publically available fingerprint database. It consists of 127163 fingerprint samples from all 10 fingers (thumb, index, middle ring and pinky) of up to 59 different subjects captured at five time-separated sessions with five samples per finger using ten different fingerprint capturing devices, including optical, capacitive and thermal ones. The whole dataset is covering a time span of five years. 41 of the volunteers participated in all five sessions. Further information regarding the data set can be found by the following link: PLUS-MSL-FP-v2 (Extended) Database Result Files and SettingsHere it is possible to download all recognition results obtained using ANSI/ISO SDK v2.21.6.3997 and IDKit SDK TODO developed
by Innovatrics as well as VeriFinger SDK 13.1 developed by Neurotechnology. The EER values are summarised in three Excel files for better
readability. Each files contains 19 different sheets representing the evaluation categories considered in the corresponding publication. NOTE: Results for Neurotechnology Verifinger v13.1 as well as IDKit SDK v. TODO are not yet in the zip archive as they need to be updated to the newest version and will be provided once finished. | |||||||
|