ICDAR 2009 Signature Verification Competition (SigComp2009)
Dr. phil. nat. Marcus Liwicki DFKI - German Research Center for Artificial Intelligence Trippstadter Str. 122 D-67663 Kaiserslautern, Germany E-mail: email@example.com Phone +49 (0) 631 20575 1200 Fax +49 (0) 631 20575 1020
Online handwriting, offline handwriting, signature, verification
The collection contains simultaneously acquired online and offline samples. The online dataset is called "NFI-online" and was processed and segmented by Louis Vuurpijl. The offline dataset is called "NFI-offline" and was scanned by Vivian Blankers from the NFI. The offline signatures were automatically segmented and visually inspected by Louis Vuurpijl. Processing of the segmented signatures was performed by Katrin Franke. The collection contains:
- authentic signatures from 100 writers
- forged signatures from 33 writers
The NLDCC-online signature collection contains in total 1953 online and 1953 offline signature files.
All writers were employees of the NFI. Each writer was asked to write his own signature on a A4 sheet of paper. Writers wrote with an inking digitizer pen on a Wacom A4-oversized tablet sampled at 200Hz at 2540 dpi resolution. The sheet of paper containing multiple boxes was fixed to the tablet.
All writers wrote inside or sometimes across the boxes, which were visible through the (blank) A4 paper. For the forgeries, each writer was asked to write a certain number of skilled forgeries on a A4 sheet of paper, maintaining the same procedure as described above.
This layout enabled relatively straight-forward segmentation in both the offline and online domains.
Writers were instructed to (i) practice in forging authentic signatures as much as they found necessary. Forgers were requested to do a free hand simulation of the authentic signatures, using a black inking pen on an A4 paper sheet on top of the tablet (ii) write each signature in the corresponding boxes. During the forging process they were allowed to use a sheet with the authentic signature as an example.
During all recordings, an experimenter was present, guiding the writers in the data acquisition process. The task of the experimenter was to collect writer details, fix each A4 sheet of paper properly, and start recording the online samples by pressing a button.
The NFI-online signature collection was recorded using:
- Wacom Intuos2 A4-oversized tablet (XD-0912-R)
- Wacom Intuos black inking pen (XP-110-00A)
- Oasis software version 8.56 (Kiko software)
- Processor: VIA C3, 800 Mhz. (http://www.kikosoft.com/products/hardware.php)
- Memory: 128 MB.
- Storage: 256 MB Flash card.
- V.L. Blankers, C.E. van den Heuvel, K. Franke, L. Vuurpijl (et al), "The ICDAR 2009 Signature Verification Competition", in Proceedings of the 2009 10th International Conference on Document Analysis and Recognition, pp. 1403-1407, 2009
- L.C. Alewijnse, C.E. van den Heuvel, R.D. Stoel, and K. Franke, "Analysis of Signature Complexity", Proceedings of the 14th Conference of the International Graphonomics Society (submitted January 2009)
- "Ink Markup Language (InkML)", W3C Working Draft 23 October 2006, http://www.w3.org/TR/InkML/#orientation.
Note: the password for opening the zip files is "I hereby accept the SigComp 2009 disclaimer." (without the double quotes).
- Disclaimer (11 Kb)
- Evaluation set (20 Mb)
- Training set (117 Mb)
This page is editable only by TC11 Officers .