FACIAL RECOGNITION BASED ELECTRONIC VOTING AND REAL-TIME RESULT COLLATION FRAMEWORK

ABSTRACT

Presently, the electoral voting process in Nigeria, involves the biometric enrolment of voter’s fingerprints, waiting to collect voter’s cards, then queuing up on the day of elections to cast their votes using ballot papers and indelible ink for validation and elimination of multiple vote casting. At the end of the vote casting exercise, the counting of the ballots begins manually, this can take a couple of days to complete, before the election results are delivered, during this time a lot of manipulation and error can be made. Multiple mobile applications were designed to acquire the photos and biodata of voters for the purpose of voter enrolment and verification of voters during elections, and a web framework was designed to facilitate the management of voter records. To make the voter enrolment and verification process reliable, facial recognition technology was used to pair each enrolled face to a unique identifier. The use of facial recognition eliminated duplicate voter registrations by a long short, while the use of Electronic Queues and Data Processing Algorithms via the web backend through Laravel PHP (Hypertext Preprocessor) Framework added the required speed needed to deliver election results on the same day of the election. Compared to existing voting systems which use fingerprints to validate voters at the voting ground, this project enables remote voting whereby voters can cast their votes securely over the internet through the mobile application created for the casting votes. It also resulted in speedy registration of voters because remote registration allows voters to be enrolled from anywhere in the country through the mobile application created for voter enrolment, due to this fact, resources no longer needed to be deployed via transport to any voter registration centre, as it is all done electronically.