Please use this identifier to cite or link to this item:
https://repositori.uma.ac.id/handle/123456789/27357
Title: | Klasifikasi Tumbuhan Obat Berdasarkan Citra Daun Menggunakan Algoritma CNN |
Other Titles: | Classification of Medicinal Plants Based on Leaf Images Using CNN Algorithm |
Authors: | Sinaga, Nicolas Novelico |
metadata.dc.contributor.advisor: | Sembiring, Arnes |
Keywords: | Medicinal Plant Classification;Leaf Image;CNN;Deep Learning;MobileNetV2;Klasifikasi Tumbuhan Obat;Citra Daun |
Issue Date: | May-2025 |
Publisher: | Universitas Medan Area |
Series/Report no.: | NPM;208160017 |
Abstract: | Penelitian ini bertujuan untuk mengklasifikasikan berbagai jenis tanaman obat berdasarkan citra daun dengan memanfaatkan algoritma Convolutional Neural Network (CNN). Model yang digunakan adalah arsitektur MobileNetV2 karena kemampuannya dalam menyeimbangkan akurasi dan efisiensi komputasi. Dataset citra daun dibagi menjadi data latih dan validasi, kemudian diproses melalui beberapa tahap,seperti augmentasi, fine-tunning, dan regularisasi. Hasil evaluasi menunjukkan bahwa model mencapai akurasi validasi tertinggi sebesar 98,43%, membuktikan bahwa pendekatan ini efektif dalam mengidentifikasi jenis tanaman obat. This study aims to classify various types of medicinal plants based on leaf images by utilizing the Convolutional Neural Network (CNN) algorithm. The model used is the MobileNetV2 architecture because of its ability to balance accuracy and computinal efficiency. The leaf images dataset is divided into training and validation data, then processed through several stages such as augmentation, fine-tunning, and regularization. The evaluation results show that the model successfully achieved the highest validation accuracy of 98,43%, proving that this approach is effective in identifying types of medicinal plants. |
Description: | 12 Halaman |
URI: | https://repositori.uma.ac.id/handle/123456789/27357 |
Appears in Collections: | SP - Informatic Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
208160017 - Nicolas Novelico Sinaga - Fulltext.pdf | Fulltext | 831.13 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.