Comparing data mining classifiers for grading raisins based on visual features

Abstract

In this study, quality grading of raisins using image processing and data mining based classifiers was investigated. Images from four different classes of raisins (green, green with tail, black, and black with tail) were acquired using a color CCD camera. After pre-processing and segmentation of images, 44 features including 36 color and eight shape features were extracted. Correlation-based feature selection was used to select best features for grading the raisins. Seven features were found superior. To classify raisins, four different data mining-based techniques including artificial neural networks (ANNs), support vector machines (SVMs), decision trees (DTs) and Bayesian networks (BNs) were investigated. Results of validation stage showed ANN with 7-6-4 topology had the highest classification accuracy, 96.33%. After ANN, SVM with polynomial kernel function (95.67%), DT with J48 algorithm (94.67%) and BN with simulated annealing learning (94.33%) had higher accuracy, respectively. Results of this research can be adapted for developing an efficient raisin sorting system.