Vote identification and integrity of ballot in paper-based e-voting system

Abstract

The paper-based e-voting system still utilises ballot and then followed by computer-based voting tabulation. The previous research does not consider the integrity of the ballot paper and the vote on top of it. In this paper, we propose to scan and store the ballot into an image file. We run a series of image processing techniques to identify the vote on the ballot image which is subsequently used in the recapitulation. We then design an automatic identification of user's vote on the ballot and protecting its integrity using Whirlpool hash algorithm. The voting ballot from an image is recognised by performing the absolute difference between the image and the provided template. The parties participating in the general elections and their candidates are stored in separated files. Based on the intensive experiments, the proposed system successfully recognises the vote with the highest accuracy at 85.71%.