Contour Based and Region Based Feature Extraction for Classification of Tubers Based on Leaf Image

Abstract

This study aims to classify the types of tubers. The amount of data used in this research is 411, which is divided into training data and testing data. The number of training data is 317, which is divided into 3 classes, namely cassava class 116, taro class 86, and uwi class 115. Total testing data used is 148, which is divided into 3 classes, namely cassava class 40, taro class 21, and uwi class 33. This research begins with segmenting the image using Otsu thresholding, then features extraction using contour-based and region-based methods. The last step is classification using K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) methods. This study provides the highest accuracy of 93% with the KNN algorithm and a neighboring value of 1 with contour-based feature extraction.