EVMFFR: Electronic Voting Machine with Fingerprint and Facial Recognition

An Internet of Things (IoT) based Electronic Voting Machine with Fingerprint and Facial Recognition (EVMFFR) with duplicate voting avoidance is proposed, designed and implemented. The designed EVMFFR system uses dual verification techniques with better results comparing to the existing single verification systems. Using raspberry, fingerprint and facial recognition methods are integrated and used for the identification of voter in the process of election. A biometric template is developed using the rectangular Haar transform during the registration of a new voter which is then used for verification and validation. The features in the developed biometric template is compared with the Eigen vectors of captured biometric of the voters using cascaded machine learning techniques such as Generalized Principal Component Analysis (GPCA) and C-Nearest Neighbor (CNN). The results of the proposed EVMFFR system shows that the integrated design performs better than the existing EVMs. The proposed EVMFFR system performs with an accuracy of more than 90% under day light for facial recognition at comparatively low cost per EVMFFR unit.