Experimental Accuracy Analysis of Multi-Biometric Authentication System Using Machine Learning Techniques

Biometrics is a set of best computerized methods utilized for recognition purpose including analyzing physical characteristics of individuals and statistically measuring them. In real world applications, fingerprint and face images are generally utilized for recognition. This research seeks to evaluate the various studies proposed to classify real and false fingerprints and faces using machine learning algorithms, as well as to evaluate various schemes. This review highlights research challenges and provides a new perspective on how to address these challenges in future. The results show that the classifiers that extract the most features are Catboost classifier, Random forest classifier and convolutional neural network. convolutional neural network models were the saturated models in biometric authentication.