Analisis Perbandingan Metode Convolutional Neural Network (CNN) dan MobileNet dalam Klasifikasi Penyakit Daun Padi

Turahman, Tazkira
Hasmin, Erfan
Aryasa, Komang

Jurnal JTIK (Jurnal Teknologi Informasi dan Komunikasi)
Vol 9 No 1 (2025): JANUARI-MARET 2025

This study aims to compare the effectiveness of Convolutional Neural Networks (CNN) and MobileNet in classifying rice leaf diseases (Oryza sativa), such as bacterial blight, brown spot, and leaf smut. The use of a dataset from Kaggle facilitates the performance evaluation of both models. The results show that MobileNet achieved a higher accuracy of 94.79% in just 10 epochs, while CNN reached 90.24% after 150 epochs. MobileNet’s efficiency in terms of training time and performance is superior to CNN. This study recommends using MobileNet for similar applications and further research with an expanded dataset and other deep learning methods.