GUC Light Field Visible Spectrum Iris Artefact Database (GUC-LF-VIAr-DB)

 

GUC light field visible spectrum iris artifact database is composed in the context of the EU FP7 FIDELITY project. project. The database collection is carried out during the period of November 2012 till April 2014 at Gjøvik University College (GUC), Norway. The database is collected using commercially available light field camera from Lytro to capture Normal (or real) iris samples and Canon 550D DSLR camera to capture corresponding high quality samples that are used to generate the artifact samples.

The whole database is composed of 52 subjects with 104 unique eye patterns. Since normal (or real) samples are captured using both light field and canon DSLR camera, this database has 520 (= 104 X 5 samples/eye) high quality samples and 4327 light field samples. The artifacts are generated by simulating 5 different kinds of attacks that ensued in 7607 samples (artifact).

Regarding the protocols for evaluation, baseline performance and more statistics refer the following publication.

R. Raghavendra, C. Busch, ‘Presentation Attack Detection on Visible Iris Recognition by Exploring Inherent Characteristics of Light Field Camera’, proceedings of International Joint Conference on Biometrics (IJCB), Clearwater, USA, 2014.

For database contact:
Dr. Raghavendra Ramachandra (raghavendra.ramachandra@ntnu.no)

Expected date of release: November 2014.