Ballot Image Test Data: RPI Synthetic Ballots


Obtaining access to hand-marked ballots created by voters for use in real elections appears to be problematic due to various legal constraints. As part of our work on the PERFECT project, we are conducting various data collection activities that will give us access to similar sorts of markings created in less restrictive contexts. We plan to make these data freely available as a service to the research community.

The ballot images below were synthesized by Anne Miller, a graduate student at Rensselaer Polytechnic Institute, using software that was written by Dan Lopresti. Briefly, this collection includes:
  1. synthetic and scanned synthetic ballots;
  2. four kinds of "noise": size, displacement, rotation, and contrast;
  3. two scans of each with different scanner settings;
  4. a blank ballot scanned with the same settings;
  5. a Kodak gray wedge scanned with the same settings as well as its spec sheet;
  6. an Excel workbook with separate sheets for each mark configuration.

Button Minnesota sample ballot. Original PDF document.
        (PDF document, 44kb)

Button Minnesota sample ballot. Original PDF document converted to TIFF and then cleaned up to remove illustrative artifacts. (Note: this version was not used to synthesize the variations listed below.)
        (TIF image, 599kb)

Button Minnesota sample ballot. Original PDF document converted to TIFF.
        (TIF image, 818kb)

Button Minnesota blank sample ballot scanned at 200 dpi.
        (TIF image, 7,788kb)

Button Minnesota blank sample ballot scanned at 200 dpi, saturated at gray-level 255.
        (TIF image, 7,788kb)

Button Minnesota sample ballot. Synthesized PDF document converted to TIFF,
marks in nominal position.
        (TIF image, 515kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
marks in nominal position.
        (TIF image, 7,788kb)

Button Minnesota sample ballot. Synthesized PDF document converted to TIFF,
marks shifted from nominal position.
        (TIF image, 514kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
marks shifted from nominal position.
        (TIF image, 7,788kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
saturated at gray-level 255, marks shifted from nominal position.
        (TIF image, 7,788kb)

Button Minnesota sample ballot. Synthesized PDF document converted to TIFF,
marks synthesized at four gray level levels.
        (TIF image, 514kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
marks synthesized at four gray level levels.
        (TIF image, 7,788kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
saturated at gray-level 255, marks synthesized at four gray level levels.
        (TIF image, 7,788kb)

Button Minnesota sample ballot. Synthesized PDF document converted to TIFF,
marks rotated +- 0, 15, 30, 45 degrees.
        (TIF image, 528kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
marks rotated +- 0, 15, 30, 45 degrees.
        (TIF image, 7,788kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
saturated at gray-level 255, marks rotated +- 0, 15, 30, 45 degrees.
        (TIF image, 7,788kb)

Button Minnesota sample ballot. Synthesized PDF document converted to TIFF,
mark size scaled 50%, 75%, 100%, 125%.
        (TIF image, 521kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
mark size scaled 50%, 75%, 100%, 125%.
        (TIF image, 7,788kb)

Button Minnesota synthesized sample ballot scanned at 200 dpi,
saturated at gray-level 255, mark size scaled 50%, 75%, 100%, 125%.
        (TIF image, 7,788kb)

Button Users’ guide and density values of steps for Kodak calibration chart.
        (PDF document, 475kb)

ButtonKodakTest
        (TIF image, 7,786kb)

Button Kodak reflectance calibration chart scanned at normal setting for ballots.
        (TIF image, 7,786kb)

Button Details of mark synthesizer specification for each ballot on separate worksheets:
  • listed by Office and Candidate column by column for Minnesota Sample Ballot;
  • legend for Mark type on worksheet "Legend;"
  • xm, ym indicate the nominal positions of the center of the target ovals in pixels at 200 dpi;
  • mark displacements in pixels relative to target locations;
  • last column is instruction to ballot mark synthesizer program generated from specifications.
        (MS Excel file, 163kb)



PERFECT is an acronym that stands for "Paper and Electronic Records for Elections: Cultivating Trust." PERFECT is a multidisciplinary research effort aimed at studying the reliable processing of paper ballots and other hardcopy election records. Participating institutions include Lehigh University, Boise State University, Muhlenberg College, and Rensselaer Polytechnic Institute. Click here to return to the PERFECT homepage.
NSF logo
 PERFECT is funded in part by the National Science Foundation under award numbers NSF-0716368, NSF-0716393, NSF-0716647, NSF-0716543. Any opinions, findings, and conclusions or recommendations expressed on this website are the investigators' and do not necessarily reflect those of the National Science Foundation.