Bell Pepper Leaf Disease Classification Using Fine-Tuned Transfer Learning

This project is the result of my Thesis on a Computer Science Master's Degree at Nusa Mandiri University in 2021. The aim is to develop a CNN model with Transfer Learning to predict bell pepper leaf disease by images using image processing.

Notebook

The code for this project is contained in the bell_pepper_leaf_desease_classification.ipynb Jupyter notebook file, which can be viewed and interactively run on platforms like Google Colab.

Paper

The project has been write into paper and has been published in the Journal of Electronics and Telecommunications / Jurnal Elektronika dan Telekomunikasi (JET) with Grade 2 accreditation on August 31, 2023, Indonesia(🇮🇩).

Abstract

Leaf diseases of plants are common worldwide. Using image processing, farmers could spot diseases in pepper plants more rapidly and get advice from plant disease experts. In this paper, researchers developed a Transfer Learning classification model for bell pepper leaf disease, with the Transfer Learning model trained on images of healthy and diseased bell pepper leaves. Classification of healthy and diseased bell pepper leaves has been carried out, and fine-tuned Transfer Learning has been applied using several pre-trained CNN models. To achieve the best outcome, four pre-trained models, including MobileNet, VGG16, ResNetV250, and DenseNet121, and three Fully Connected (FC) layer architectures were tested. The Fully Connected (FC) layer with four Transfer Learning architectures achieved the best accuracy value of 99.33% on DenseNet121 architecture with one layer and Cohen’s Kappa value of 0.9865.

DOI : http://dx.doi.org/10.55981/jet.546

Dataset

This research uses images on a public dataset obtained from Tairu Oluwafemi Emmanuel published in Kaggle entitled PlantVillage Dataset.

The images available on the dataset are tomato leaf, potato leaf and paprika leaf. However, we're only focused on using the datasets, i.e. just bell pepper leaf images. The dataset used have 2 classes, Bacterial (a) and Healthy (b).

Image: Sample Dataset

Cited By IEEE Paper :

1. GSAtt-CMNetV3: Pepper Leaf Disease Classification using Osprey Optimization
2. Analysis of Pre-Trained CNN Models for Pepper and Potato Leaf Disease Prediction

Implementation App

This repository contains web-based applications built using python, flask and some additional libraries listed in the requirements.

Back-End

• Python 3.8.5

Requirements

• Werkzeug 2.0.1
• Flask 2.0.1
• Numpy 1.19.5
• Keras 2.4.3
• Tensorflow 2.14.0
• Gevent 21.1.2
• Pillow 8.2.0
• H5py 2.10.0
• Gunicorn 20.1.0

Screenshot

Home Page
Prediction Page

Install App

Choose your local folder destination, open the terminal, and do these steps :

git clone https://github.com/yuris60/bell_pepper_leaf_desease_classification.git

pip install -r requirements.txt

python app.py

Open Browser and type http://localhost:5000

Cite our paper

We would be very grateful if you could cite our work in your research.

BibteX Format

@ARTICLE{Akhalifi2023-ym,
  title     = "Bell pepper leaf disease classification using fine-tuned Transfer Learning",
  author    = "Akhalifi, Yuris and Subekti, Agus",
  journal   = "J. Elektron. Dan Telekomun.",
  publisher = "National Research and Innovation Agency",
  volume    =  23,
  number    =  1,
  pages     = "55",
  month     =  aug,
  year      =  2023
}

IEEE Format

Y. Akhalifi,  and A. Subekti, "Bell Pepper Leaf Disease Classification Using Fine-Tuned Transfer Learning," Jurnal Elektronika dan Telekomunikasi, vol. 23, no. 1, pp. 55-61, Aug. 2023. doi: 10.55981/jet.546

Thanks To

Thank you very much to Dr. Agus Subekti, M.Kom for his guidance to me during the preparation and development of the project and publication papers. Without him, this project couldn't have been completed as it should be.