BI-DIRECTIONAL CLASSIFICATION OF ROMAN PERIOD COINS BY DEEP LEARNING METHODS

In this study, the problem of classification coins, which have historical importance and can only be distinguished by experts, is discussed with pre-learning deep learning algorithms. solution problem, RRC-60 dataset, consists images coins used in Roman Republic period, was used. Xception, MobileNetV3-L, EfficientNetB0 DenseNet201 models were trained using on both sides data set. As a result training, best values, Precision, Recall F1-Score metrics MobileNetV3-L model 98.2%, 96.8%, 97.5%, respectively, test accuracy 95.2%