Classification of Marble Types Using Machine Learning Techniques

Natural stones are one of the indispensable elements people from shelter to weapons. Among these stone types, marbles and marble-derived products among objects that always prefer, bathroom kitchen, garden design small decorative home decorations. While named according regions where they extracted, their types qualities classified based on observation by who qualified as experts in this field. This classification, which is made observation, carries risks economic terms, increases workload a difficult process with high error rate. These processes need fast, easy highly accurate digital transformation. In study, feature extraction was done using deep learning species classification marbles. The extracted features were machine techniques. As result application data set consisting 3703 marble natural images belonging 28 different species, success 99.7% obtained DenseNet model K-Nearest Neighbor method.