PERFECT Project: Homogeneous Class Display

Manually processing large quantities of ballot images in an audit or recount is a tedious and error-prone activity. Another area of our investigation is to explore approaches to facilitate detecting errors in machine or human interpretation of ballot images.

Our idea designed to verify automated mark sense counts is inspired by long-established methods of OCR verification, homogeneous class display (HCD). Here the isolated images where a mark was registered as a vote are grouped for display. The images where a mark was not registered are similarly grouped. The positive marks are displayed in groups of, say, 64 (8x8) simultaneously, and the human operator
is asked to note any discrepancies. The negative instances are displayed in similar groupings, and again the operator will notate any perceived discrepancies. This is depicted in the BallotTool screen snapshot below, where several potential recognition errors are evident on quick examination.

In a real application, spurious marks can also be introduced to assess operator error rate. (In some OCR service bureaus, the operators are notified of the presence of artificial errors in order to keep them alert.) HCD is, in principle, a very fast and convenient method of verifying automated counts, but it cannot be accepted without experimental evidence of its accuracy.

Homogeneous class display
(a) Homogeneous class display.

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.