Comparative Analysis of MobileNetV2 and EfficientNetB0 for Face and Fingerprint Recognition in Machine Learning Enhanced Access Control

Abstract

This paper presents a comparative analysis of MobileNetV2 and EfficientNetB0 for face and fingerprint recognition in amachine learning-enhanced access control system. The study was developed within an edge computing security frameworkwhere biometric verification was integrated with contextual risk analysis and Chinese Wall policy enforcement.MobileNetV2 was adopted for facial recognition because of its lightweight convolutional architecture and suitability forefficient image classification. EfficientNetB0 was adopted for fingerprint recognition because of its compound scalingstrategy and strong feature extraction capability. The models were trained and evaluated using locally captured real-userbiometric samples organized into ten identity classes. MobileNetV2 achieved approximately 97% target accuracy for facerecognition, while EfficientNetB0 achieved approximately 93.36% accuracy for fingerprint recognition. Although the facemodel produced higher reported classification accuracy, the fingerprint model provided strong operational stability as aphysiological verifier. The findings support the use of both models in a bimodal biometric access-control framework.Keywords- MobileNetV2, EfficientNetB0, face recognition, fingerprint recognition, machine learning, deep learning,bimodal biometrics, access control, edge computing, biometric authentication.