MASIVE: A Dataset of Multi-impression Fingerprint Samples used for the Investigation of Verification Failures in an African Election Setting

The dataset of fingerprints known as MASIVE (derived from the words Multi-impression, African, Samples, Investigation, Verification, and Election) comprises fingerprint images of 289 Nigerian participants and corresponding metadata that were collected in typical election settlings in Nigeria. The purpose of collection is to investigate the likely causes of fingerprint verification failures that were experienced by a reasonable proportion of the electorate during the Country’s previous national elections, and proffer possible solutions that could potentially help improve the verification rates at Her future elections. These solutions might also be beneficial to other Sub-Saharan African countries that have adopted the use of fingerprint verifications for their elections. However, the dataset could be used for other fingerprint recognition related research as well. A total of 12403 fingerprint images of both the index fingers and thumbs of the participants were collected. The fingerprints were grouped based on 6 different impressions as might be obtainable in the operating settings of an election. The dataset has participants from age groups 18 years (inclusive) to 99 years whose fingerprint images were collected using a light emitting sensor (LES) scanner.
Besides grouping the fingerprint by impressions, the dataset includes two CSV files that contain the following metadata.
The age group, gender, and occupation of each participant.
Presence or absence of scars on the fingertips, from physical observation.
Record of verification related issues experienced by participants during past elections and at other times (e.g., biometric capturing for national identification).
Temperature and humidity at the times of capturing.
While this dataset could potentially be contributory to strengthening the democracies of Sub-Saharan Africa countries by contributing towards improved electoral process, it is also a valuable addition to the fingerprint biometric research community at large.

Requests to access the data are made directly to the repository, via the [email protected] email. Requests will be reviewed by a team composed of academic research, and governance staff who will consider whether the proposed reuse of the data aligns with the commitments stated in the participant information sheet (PIS) issued during data collection. End users are required to assent to an End User Licence Agreement that stipulates the data can only be used for scientific research similar to the research described in the PIS and that is in the general interest of the public. All traces of the data in the End User's custody must be destroyed within two years of first access.

Applications for access to the dataset will be considered from bona fide researchers with a direct affiliation to a research organisation only. Applicants will be asked to provide documentary evidence that ethical approval has been granted by their organisation prior to obtaining access. The organisation must be operating rigorous research governance processes. Further conditions are outlined within the University of Dundee End User Licence Agreement this is available to download under the Access Dataset field.

Data currently under embargo with intent to publish research findings, for more information contact [email protected]