Aug 2026· International Research Journal of Computer Science· Vol 13, pp. 622· 0 citations· 30 references
TL;DR
This work presents a convolutional neural network based on AlexNet that is lightweight and field-aware, able to perform inference faster without sacrificing accuracy, making it suitable for real-time implementation in low resource agricultural environments.
Abstract
Tomato leaf diseases must be identified early and accurately in order to reduce output loss and advance sustainable agriculture. Deep learning models have shown encouraging results in the identification of plant diseases, but their high processing requirements and inability to adjust to field-specific limitations sometimes make it difficult to implement them in real-world applications. For the purpose of accurately and efficiently classifying tomato leaf diseases, we present a convolutional neural network based on AlexNet that is lightweight and field-aware. Our proposed model is designed with less complexity than traditional architectures and is able to perform inference faster without sacrificing accuracy, making it suitable for real-time implementation in low resource agricultural environments. The algorithm was trained and tested on a dataset of 7704 augmented images of tomato leaves from 7 different disease categories. The Lightweight AlexNet achieved accuracy comparable to VGG variants and outperformed deeper models such as ResNet (96.81%) with lower parameter overhead. It was trained from scratch using TensorFlow & Keras on 100×100 pixel inputs, achieving a training accuracy of 99.15% and a validation accuracy of 99.74%. Moreover, an effective structure of the model enables the installation on edge devices, which offers a scalable precision farming solution. Our work helps to bridge the gap between deep learning research and real-world application in agriculture, allowing the development of real-field, resource-efficient disease detection systems.
A modified AlexNet architecture for classifying field-captured tomato leaf images into seven disease categories was developed and generally focused on symptom-bearing leaf regions, whereas target spot was the most difficult category to classify.
Debabrat Bharali, Kanak C. Bora, Rashel Sarkar et al.· Journal of Scientific Resear...· 0 citations
Agricultural production faces significant annual losses due to plant diseases, with economic impacts exceeding 40 million dollars and contributing to acute hunger affecting over 281.6 million people in 2023. The timely and accurate identification of plant diseases through leaf image analysis is crucial to mitigate these losses and ensure global food security. This study proposes a lightweight Convolutional Neural Network model, inspired by the MobileNet architecture, designed to classify various plant leaf diseases efficiently. Leveraging a publicly available dataset, this research focuses on developing a model that balances high performance with computational efficiency, making it suitable for real-world applications in resource-constrained environments. The proposed model, named LDPNet, achieved an outstanding accuracy of 99.62%, alongside precision, recall, F1-Score, and AUC metrics of 99.16%, 99.08%, 99.11%, and 99.99%, respectively. A comprehensive comparative analysis was conducted against MobileNetV2 and a reference model from previous research, highlighting the superior performance of LDPNet in terms of both accuracy and efficiency. The results demonstrate that the proposed architecture not only maintains high classification performance but also significantly reduces the number of parameters, making it a practical and scalable solution for plant disease identification. This study contributes to the growing field of agricultural technology by providing a robust, lightweight, and efficient tool for early disease detection, with the potential to enhance crop management and reduce economic losses in agriculture.
Roney Nogueira De Sousa, Saulo Anderson Freitas De Oliveria, P. Rebouças· Journal of the Brazilian Com...· 0 citations
This research presents the evaluation of plant leaf disease detection system based on a convolutional neural network (CNN) optimized with different nature-inspired algorithms, and indicates that application of the grey wolf optimizer (GWO) for tuning key hyperparameters of the model produced the best results.
Agricultural productivity is also at risk of plant diseases, especially in areas where the lack of experts makes it hard to diagnose the disease timely. Conventional manual inspection is slow, subjective and impractical when dealing with large scale monitoring. To mitigate this issue, this paper introduces a lightweight deep learning architecture that will be used to classify plant leaf diseases in real-time. The given solution makes use of the idea of transfer learning based on MobileNetV3 to ensure a high level of classification accuracy and a low complexity of computations. The Plant Village data, which is publicly available, and consists of various leaf images of many different crops and disease classes, was used to carry out the experiments. Normal preprocessing and data augmentation methods were used to enhance the generalization of the model. The proposed model had an overall accuracy of 94% with a high level of precision and recall by all the classes. Moreover, the lightweight architecture can be quickly inferred and deployed to a mobile and edge device, which makes it very appropriate. The findings show that an effective deep learning model can deliver valid and real-time diagnosis of a disease, which can be used to support practical agriculture and sustainable crop management.
R. Balamanigandan, A. Jenifer· 2026 4th International Confe...· 0 citations
The FRCNN-FCMWeight model provides strong support for agricultural use cases and improves overall model generalization and enhances training robustness and identifies diseased regions effectively under highly variable field conditions.
Shashank Yadav, A. S. Tewari· Journal of Crop Health· 0 citations
