CBIR for Herbs Root Using Color Histogram and GLCM Based on K-Nearest Neighbor

摘要

Ginger, kencur, kunyit, temulawak, and temu hitam are herbs root that is widely used as a spice and herbal medicine in Indonesia. Visually the shape and color of herbs root are quite similar so that the common people difficult distinguish. The problem can be solved with the help of Content-Based Image Retrieval (CBIR). There are three main modules in CBIR namely pre-processing, feature extraction and identification. The study proposed the Color Histogram and Gray Level Co-occurrence Matrix (GLCM) as feature extraction to identify these types of herbal remedies. While the identification stage used K-Nearest Neighbor (KNN) and the calculation of the distance closest to the Euclidean distance. The amount of data used as training data is 250 images, while data testing used 25 images. The results of this study the system can recognize the image of herbs root in accordance with the type with the highest accuracy obtained.