International Journal on Data Science and Technology 2019; 5(3): 57-65 http://www.sciencepublishinggroup.com/j/ijdst doi: 10.11648/j.ijdst.20190503.11 ISSN: 2472-2200 (Print); ISSN: 2472-2235 (Online) Classification of Marble Using Image Processing Fisha Haileslassie 1, * , Adane Leta 2 , Gizatie Desalegn 1 , Meles Kalayu 3 1 Department of Computer Science, Faculty of Technology, Debre Tabor University, Debre Tabor, Ethiopia 2 Colleage of Informatics, University of Gondar, Gondar, Ethiopia 3 Department Information Technology, Raya University, Maichew, Ethiopia Email address: * Corresponding author To cite this article: Fisha Haileslassie, Adane Leta, Gizatie Desalegn, Meles Kalayu. Classification of Marble Using Image Processing. International Journal on Data Science and Technology. Vol. 5, No. 3, 2019, pp. 57-65. doi: 10.11648/j.ijdst.20190503.11 Received: November 13, 2019; Accepted: December 18, 2019; Published: December 24, 2019 Abstract: Classification of marble image according to usage purpose and quality is an important procedure for export. Discrimination between marble varieties is a difficult task during selection, since it requires trainings and experience. Therefore, the development of automatic prediction model based on image processing is a potential application area to support experts across the world. In this study an attempt has been made to develop marble variety classification model by comparing color, texture and ensemble of color and texture. In view of this, a digital image processing technique based on combined texture and color features have been explored good classification performance to classify varieties of marble image. On the average 60 images were taken from each of the three marble varieties (Grade A, Grade B, Grade C). The total number of images taken was 180. For the classification model we applied image preprocessing techniques; image acquisition, image conversion, noise removal, image enhancement, edge detection and image binarization. For texture extraction gray level co- occurrence matrix, for color extraction color histogram was applied. For classification five textures and six color features were extracted from each marble image. To build the classification models for prediction of marble varieties, K-Nearest Neighbors (KNN), Artificial Neural Network (ANN) are investigated. Based on experimental results, ANN outperforms KNN. Quantitatively, an average accuracy of 83.3% and 93.7% is achieved KNN and ANN respectively for Grade A, Grade B, Grade C varieties with the combined feature sets of color and texture. This shows an encouraging result to design an applicable marble classification model. Marble fractured and vines of the images affect greatly the performance of the classifier and hence they are the future research direction that needs an investigation of generic noise removal and feature extraction techniques. Keywords: Classification of Marbles, Feature Extraction, Artificial Neural Network Classifier, K-nearest Neighbor, Sobel Edge Detector 1. Introduction Image processing deals with manipulation of digital images through a digital computer for analyzing image data; it is the realization of almost all the services of the human visual system through computers [1]. It is a type of signal dispensation, which outputs an image or characteristics associated with that image. The main objectives of image processing are to create more suitable images for people to observe, identify and, computers can automatically recognize and understand images [1]. Image-processing technology, has rapidly developed since the 1960s and 1970s [1, 2]. It is basically the technique of manipulating and improving gray-scale images by using mathematical functions. Image-processing system contains three fundamental elements: an image-acquisition element, an image-processing element, and an image-display element. Image acquisition is the process of obtaining a digital image of object in real world using camera or scanner [1]. This image processing element are image enhancement, image analysis, and image-coding [1, 2]. For human viewing, image enhancement improves the quality and clarity of images. Image analysis involves calculations on a final image to produce numerical results. Image analysis in general is to extract, from the very large amount of data in an image, that small set of measurements containing the information of