Classification of Rice Leaf Diseases from Digital Images using Convolutional Neural Network

Widiasri, Monica and Siswantoro, Joko and Pratama, Putu Pramodia Suka (2026) Classification of Rice Leaf Diseases from Digital Images using Convolutional Neural Network. In: 2025 5th International Conference of Science and Information Technology in Smart Administration (ICSINTESA), 29-29 Nov. 2025, Yogyakarta.

	PDF ICSINTESA revision final manuscript 1571192720.pdf Download (980kB)
	PDF ICSINTESA MonicaW.pdf Download (1MB)

Abstract

Rice (Oryza sativa L.) is one of the main food crops in Indonesia. The Indonesian people rely on rice as their primary source of carbohydrates, so the quality and health of rice during the planting process must be considered to minimize the risk of crop failure. One of the causes of rice crop failure is the attack of pests or pathogens that attack rice. Pests or pathogens that attack rice plants cause diseases with a similar visual appearance, making it difficult for farmers to distinguish them. The purpose of this research is to create a convolutional neural network (CNN) model that is able to classify diseases on rice leaves from digital images to help farmers in distinguishing the visual characteristics of rice leaf diseases. Digital image data of five types of rice leaf diseases, namely bacterial leaf blight, blast, brown spot, leaf smut, and tungro were collected from three different sources. The CNN model used is a custom CNN model, which was evaluated with hyperparameter tuning and k-fold cross-validation. The best model was then retrained with a data ratio of 70:15:15 (training:validation:test) and obtained an accuracy metric equal to 0.9870; average precision equal to 0.9869; average recall equal to 0.9872; and average f1 score equal to 0.9870. The model was then implemented into an Android application using TensorFlow Lite. Based on the survey results, respondents agreed that the application can make it easier for farmers to classify the types of rice leaf diseases based on visual displays using digital images of rice leaf diseases.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Classification, Rice Leaf Disease, Convolutional neural network, K-Fold cross-validation, Android Application
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions:	Faculty of Engineering > Department of Informatic
Depositing User:	Monica Widiasri 61151
Date Deposited:	05 May 2026 07:01
Last Modified:	05 May 2026 07:01
URI:	http://repository.ubaya.ac.id/id/eprint/50695

Actions (login required)

View Item