Door Lock Access Control using Fusion of Face and Fingerprint Recognition

This paper proposes an application of biometric- based fusion for doorlock access to improve room security. Generally, using a biometric system, door lock access technology still uses a unimodal approach. Some products are pretty good, but the gap for improvement is still open for further development. A fusion approach can be used to create a multimodal biometric system to deal with the progress. In this research, we propose combining facial and fingerprint recognition technologies to increase the security of the system. At the score level, we used a combination of two access technologies based on multimodal bio-metrics. Face modalities were captured with a webcam camera, the haar-cascade technique was used to detect them, and the CNN dlib method was used to recognize them. The FPM10A sensor was used to capture the fingerprint modality, the characteristics were extracted using the KAZE method, and the matching was done using the Brute-Force algorithm. A comparison of the False Acceptance Rate (FAR) and False Rejection Rate (FRR) between the unimodal and multimodal biometric systems was performed to evaluate the performance of the multimodal biometric system. The resulting FAR is 10%, with an FRR of 11% and a Genuine Acceptance Rate (GAR) of 90%, according to our experiments. Thus, the system that we have created and developed is quite capable of being used as an alternative door lock access method in addition to other types of door locks.