作者: Kadir Sercan Bayram , Bülent Bolat
DOI: 10.1186/S13640-018-0274-X
关键词: Feature (computer vision) 、 Multilayer perceptron 、 Decision tree 、 Face (geometry) 、 Biometrics 、 Feature extraction 、 Artificial intelligence 、 Pattern recognition 、 Computer science 、 Probabilistic neural network 、 Support vector machine
摘要: In this work, a secure multibiometric system is proposed. Three different biometric modalities which are ear, face, and thermal face considered. The data were taken from USTC NVIE Spontaneous Database, whereas the ear collected IIT Delhi Ear Image Database. For each modality, three feature extraction methods used four classifiers (multilayer perceptron, decision tree, support vector machines, probabilistic neural network) trained by using two fusion matching score level fusion. According to results, individual biometrics better for identification problem. However, validation problem, both give false acceptance rate/false rejection rate values regarding biometrics.