Abstract:Rice plays a vital role as a primary food source for over half of the world's population, and its production is critical for global food security. Nevertheless, rice cultivation is frequently affected by various diseases that can severely decrease yield and quality. Therefore, early and accurate detection of rice diseases is necessary to prevent their spread and minimize crop losses. In this research, we explore three mobile-compatible CNN architectures, namely ShuffleNet, MobileNetV2, and EfficientNet-B0, for rice leaf disease classification. These models are selected due to their compatibility with mobile devices, as they demand less computational power and memory compared to other CNN models. To enhance the performance of the three models, we added two fully connected layers separated by a dropout layer. We used early stop creation to prevent the model from being overfiting. The results of the study showed that the best performance was achieved by the EfficientNet-B0 model with an accuracy of 99.8%. Meanwhile, MobileNetV2 and ShuffleNet only achieved accuracies of 84.21% and 66.51%, respectively. This study shows that EfficientNet-B0 when combined with the proposed layer and early stop, can produce a high-accuracy model. Keywords: rice leaf detection; green AI; smart agriculture; EfficientNet;

What problem does this paper attempt to address?

The main problems that this paper attempts to solve are to improve the accuracy and efficiency of rice leaf disease identification, and at the same time ensure that the model can run efficiently on mobile devices. Specifically, the research aims to develop a lightweight convolutional neural network (CNN) model to achieve early and accurate detection of rice leaf diseases, thereby helping farmers take timely measures to prevent the spread of diseases and reduce crop losses. ### Research Background and Problems 1. **Global Food Security**: Rice is the main food source for more than half of the world's population, and its yield is crucial to global food security. 2. **The Impact of Rice Leaf Diseases**: Rice leaf diseases can significantly reduce the yield and quality of rice. Therefore, early and accurate disease detection is crucial for preventing disease spread and minimizing crop losses. 3. **Limitations of Existing Technologies**: Although existing deep - learning models have achieved high accuracy in rice leaf disease detection, these models usually require a large amount of computing resources and are difficult to be applied in real - time on mobile devices. Therefore, it is necessary to develop lightweight and efficient models to meet the needs of practical applications. ### Research Objectives - **Improve Accuracy**: Improve the accuracy of rice leaf disease classification by optimizing the model structure and training methods. - **Reduce Computing Resource Requirements**: Select a lightweight CNN architecture suitable for mobile devices, reduce computing resource consumption, and achieve green AI. - **Achieve Real - Time Monitoring**: Ensure that the model can run efficiently on mobile devices, and support real - time monitoring and decision - making support. ### Solutions Researchers selected three lightweight CNN architectures - ShuffleNet, MobileNetV2, and EfficientNet - B0, and made improvements to them: - **Model Structure Optimization**: Added two fully - connected layers and one dropout layer on the basis of the original model to enhance the model performance and prevent overfitting. - **Training Strategy**: Use the early stopping method to prevent the model from overfitting, and adjust hyper - parameters such as the learning rate to obtain the best performance. - **Evaluation Indicators**: Evaluate the model performance through indicators such as accuracy, recall, precision, and F1 - measure. ### Experimental Results The experimental results show that the EfficientNet - B0 model performs the best, with an accuracy rate of 99.8%, while the accuracy rates of MobileNetV2 and ShuffleNet are 84.21% and 66.51% respectively. This indicates that EfficientNet - B0 combined with the proposed improvement methods can significantly reduce the computing resource requirements while maintaining high accuracy, and is suitable for real - time rice leaf disease detection on mobile devices. ### Summary This research solves the problem of large computing resource requirements of existing rice leaf disease detection models by developing a lightweight CNN model, and provides an effective solution for achieving green AI and intelligent agriculture.

Advancing Green AI: Efficient and Accurate Lightweight CNNs for Rice Leaf Disease Identification

Enhancing Rice Leaf Disease Classification: A Customized Convolutional Neural Network Approach

Towards Sustainable Agriculture: A Novel Approach for Rice Leaf Disease Detection Using dCNN and Enhanced Dataset

Deep Learning Utilization in Agriculture: Detection of Rice Plant Diseases Using an Improved CNN Model

Utilizing Convolutional Neural Networks for the Effective Classification of Rice Leaf Diseases Through a Deep Learning Approach

An Enhancement of CNN Algorithm for Rice Leaf Disease Image Classification in Mobile Applications

Rice Leaf Disease Recognition using Local Threshold Based Segmentation and Deep CNN

Automatic Recognition of Rice Leaf Diseases Using Transfer Learning

Analysis of Convolutional Neural Network-based Image Classifications: A Multi-Featured Application for Rice Leaf Disease Prediction and Recommendations for Farmers

Resource-optimized cnns for real-time rice disease detection with ARM cortex-M microprocessors

Detection and classification of plant diseases in leaves through machine learning

Identification and recognition of rice diseases and pests using convolutional neural networks

Alpha-EIOU-YOLOv8: An Improved Algorithm for Rice Leaf Disease Detection

Comparative Analysis of Transfer Learning, LeafNet, and Modified LeafNet Models for Accurate Rice Leaf Diseases Classification

A real-time approach of diagnosing rice leaf disease using deep learning-based faster R-CNN framework

Deep Learning based Leaf Disease Detection using Convolutional Neural Network

Rice Leaf Disease Classification Using Cnn

Automatic Detection of Rice Disease in Images of Various Leaf Sizes

Detection and Prediction of Rice Leaf Disease Using a Hybrid CNN-SVM Model

A novel technique predicting the rice leaf diseases using Convolutional Neural Network

Leveraging Pre-trained CNNs for Efficient Feature Extraction in Rice Leaf Disease Classification