Corn Leaf Disease Classification Optimization Using Resnet50 Architecture Utilizing Bayesian Optimization
Authors (s)
(1) * Yahya Auliya Abdillah  


        Indonesia
(2)  Kusrini Kusrini   (Universitas Amikom yogyakarta)  
        Indonesia
(*) Corresponding Author
AbstractThis research aims to optimize the classification of diseases on corn leaves using Convolutional Neural Network (CNN) architecture, ResNet50, combined with hyperparameter optimization techniques using Bayesian Optimization. The dataset used comes from Kaggle, consisting of four classes of corn leaf diseases, namely corn leaf spot, leaf rust, corn leaf blight, and healthy corn leaves. Data pre-processing was done to balance the amount of data between classes and reduce the risk of overfitting. This study tested various scenarios, including the use of the original dataset and a pre-processed dataset. The experimental results show that the use of Bayesian Optimization in hyperparameter search gives better results than manual parameter setting. The scenario with hyperparameter optimization using Bayesian Optimization technique on the pre-processed dataset shows an increase in accuracy by 5% (87.79%) compared to the scenario without optimization (82.82%). This research concludes that hyperparameter optimization techniques and proper data pre-processing can improve the performance of CNN models in corn plant disease classification, providing the potential to assist farmers in detecting diseases earlier and reducing the economic losses incurred.
|
Keywords
Convolutional Neural Network; Bayesian Optimization; Classification; Corn Leaf; Deep Learning;
Full Text: PDF
Refbacks
- There are currently no refbacks.
Copyright (c) 2025 Kusrini Kusrini

This work is licensed under a Creative Commons Attribution License (CC BY-SA 4.0)
Journal of Electrical Engineering and Computer (JEECOM)
Published by LP3M Nurul Jadid University, Indonesia, Probolinggo, East Java, Indonesia.