Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal CCIT (Creative Communication and Innovative Technology) Journal
Pagiu, Harry T.
Unknown Affiliation
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search
Journal : CCIT (Creative Communication and Innovative Technology) Journal

 1 
CNN Algorithm for Herbal Leaf Classification Using MobileNetV2 and ResNet50V2 Pagiu, Harry T.; Kasim, Anita Ahmad; Lapatta, Nouval Trezandy; Pratama, Septiano Anggun; Laila, Rahma
CCIT (Creative Communication and Innovative Technology) Journal Vol 18 No 2 (2025): CCIT JOURNAL
Publisher : Universitas Raharja

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33050/ccit.v18i2.3776

Abstract

Indonesia is home to over 30,000 types of herbal plants, with approximately 1,200 species utilized as raw materials for alternative and traditional medicine. Leaves play a crucial role in herbal medicine preparation. However, many people struggle to identify different herbal leaves due to their similar appearances, making classification difficult. Each leaf possesses unique characteristics such as shape, size, midrib, stalk, blade, and type, which can be used for differentiation. To assist in identifying herbal leaves, a classification system based on image recognition is essential. Convolutional Neural Networks (CNN) are deep learning algorithms designed for processing two-dimensional image data. Model performance can be enhanced through transfer learning, with MobileNetV2 and ResNet50V2 being widely used architectures. These pretrained models have been trained to recognize images with high accuracy. This study focuses on classifying herbal plants based on leaf shape using CNN architectures from MobileNetV2 and ResNet50V2. The evaluation results show that the MobileNetV2 architecture, with a 90%:10% data split, achieved an accuracy of 98.51%, precision of 98.92%, recall of 98.51%, and an F1-score of 98.56%. These findings indicate that CNN with transfer learning can effectively classify herbal leaves with high accuracy.
Co-Authors Anita Ahmad Kasim Laila, Rahma Nouval Trezandy Lapatta Pratama, Septiano Anggun