Implementasi Deep Learning Untuk Identifikasi Tanaman Rimpang Menggunakan Metode Convolutional Neural Network

Mahendri, Diffa Rahmanda Putra
T. Yudi Hadiwandra

Jurnal Computer Science and Information Technology (CoSciTech)
Vol 6 No 1 (2025): Jurnal Computer Science and Information Technology (CoSciTech)

Rhizome plants are spices widely used by Indonesian people as cooking ingredients or traditional medicine. These plants havesimilar appearances, making them difficult to distinguish for some people. Errors in identifying rhizome plants can lead topoisoning, allergies, or unwanted side effects. To simplify identifying these plants, a system is needed to detect and differentiatetypes of rhizome plants, which can be achieved using Convolutional Neural Networks (CNN) with the YOLO algorithm. CNN isa Machine Learning technique capable of identifying objects based on their visual features, enabling efficient differentiation ofrhizome plants. The image dataset used is divided into six classes, with a total of 700 images. Model testing produced resultswith a precision of 98%, recall of 99%, and mAP50-95 of 96%. Future research is expected to increase dataset variety to avoidoverfitting.