Classification and Analysis of Pistachio Species with Pre-Trained Deep Learning Models

Pistachio is a shelled fruit from the anacardiaceae family. The homeland of pistachio Middle East. Kirmizi pistachios and Siirt are major types grown exported in Turkey. Since prices, tastes, nutritional values these differs, type becomes important when it comes to trade. This study aims identify two pistachios, which frequently Turkey, by classifying them via convolutional neural networks. Within scope study, images were obtained through computer vision system. pre-trained dataset includes total 2148 images, 1232 916 type. Three different network models used classify images. Models trained using transfer learning method, with AlexNet VGG16 VGG19. divided as 80% training 20% test. As result performed classifications, success rates AlexNet, VGG16, VGG19 94.42%, 98.84%, 98.14%, respectively. Models’ performances evaluated sensitivity, specificity, precision, F-1 score metrics. In addition, ROC curves AUC performance evaluation. highest classification was achieved model. results reveal that methods can be successfully determination types.