Automatic amber gemstones identification by color and shape visual properties $ Saulius Sinkevicius n , Arunas Lipnickas, Kestas Rimkus Department of Automatics, Kaunas University of Technology, Studentu St. 48-320, LT-51367 Kaunas, Lithuania article info Article history: Received 10 June 2014 Received in revised form 13 September 2014 Accepted 18 September 2014 Keywords: Expert systems Image classification Image matching Supervised learning abstract This paper proposes and describes novel techniques for the amber gemstones labeling system. The amber data used in experiments are collected by amber art craft industry experts and the presented investigations were carried out in order to develop a classifier for online amber sorting application. Amber pieces are identified and labeled to one of 30 color classes or to one of 20 geometric shape classes. For identification Quadratic Discriminant Analysis, K Nearest Neighbors, Radial Basis Function, Naive Bayes, Decision Tree, and pruned Decision Tree classifiers were tested. As color descriptive features mean, standard deviation, kurtosis, and skewness calculated on amber pixels from grayscale and HSV color spaces were chosen. The best classification result with the features calculation on all the pixels of sample was 69.29% accuracy, obtained by Pruned Decision Tree classifier. In order to improve the classification results, the pixels of amber samples were grouped into predefined concentric ring segments and the accuracy rose by 10%. Then the final improvement was introduced by forming a committee of Decision Tree classifiers with Half&Half method which increased accuracy up to 81.60%. For shapes identification the Centroid Distance Function was selected as it preserves the order of landmark points. Using labeled samples the Decision Tree classifier was trained. The training of classifier was made by acquiring all possible orientations of Centroid Distance Function for each image in training set and then feeding them to Decision Tree. In the classification step all the shifted and flipped Centroid Distance Function variations of the testing sample are voting for the class using the trained Decision Tree. Experimental results have shown that the proposed technique is effective in organic shapes classification to selected geometric shapes even if there is high ambiguity between organic shapes and 72.10% accuracy was acquired. Both proposed classifiers can be used in real time application independently or in combination. & 2014 Elsevier Ltd. All rights reserved. 1. Introduction Baltic amber gemstones are still mined in present time and used as decorative component for jeweler, souvenirs, or art paintings. The smallest pieces of amber, combining their color tones, transparency, variegation, shape, size and other interface features is mostly used by art crafters. At this time sorting of amber pieces according to the color and the similarity to geo- metric shapes is complicated and time consuming process, manly performed by “the eye of human being”. The presorted gemstones may be further used for automated craft making. The best solution for amber classification by color and similarity to geometric shapes is the implementation of fully automated industrial sorting line based on machine vision. Other researchers acquired good results by extracting visual features and used them for classification and sorting task into small number of categories. For the surface visual evaluation the first and second order statistical features were used (Rozman et al., 2006). Such sorters usually are used in food industry (Pearson et al., 2012) by grading fruits, garbage recycling (Shahrani et al., 2010), etc. Many of them are based on optical properties of object surface but uses different kind of sensors like CCD cameras, spectroscopy (O’Farrell et al., 2005), stereo vision, infrared light, and so on. Optical sensors are able acquire color, shape, texture, and other optical features and in many cases it is a multiclass (Aly, 2005; Ramanan et al., 2011) identification problem. Optical properties depends on lighting condition, so isolating objects from environment and implementing artificial lighting source may be Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/engappai Engineering Applications of Artificial Intelligence http://dx.doi.org/10.1016/j.engappai.2014.09.011 0952-1976/& 2014 Elsevier Ltd. All rights reserved. ☆ The material in this paper was partially presented at the 7th IEEE Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS’2013), September 12-14, 2013, Berlin, Germany. n Corresponding author. Tel.: þ370 652 60978. E-mail addresses: saulsink@gmail.com (S. Sinkevicius), arunas.lipnickas@ktu.lt (A. Lipnickas), kestas.rimkus@gmail.com (K. Rimkus). Engineering Applications of Artificial Intelligence 37 (2015) 258–267