Jul 2026· Journal of Scientific Research and Reports· 0 citations
TL;DR
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.
Abstract
Timely and accurate identification of tomato leaf diseases is important for reducing crop losses and supporting sustainable crop management. This study developed a modified AlexNet architecture for classifying field-captured tomato leaf images into seven disease categories. A custom dataset obtained under variable backgrounds and lighting conditions was expanded through rotation, horizontal and vertical flipping, and zooming to reduce class imbalance and improve model generalisation. The proposed architecture comprised three convolutional layers with 3 × 3 filters, reduced fully connected layers, and dropout regularisation to balance predictive performance and computational efficiency. The model was implemented using TensorFlow and Keras and trained for 30 epochs. Its performance was evaluated using accuracy and loss curves, a confusion matrix, receiver operating characteristic curves, area under the curve values, class-wise precision, recall and F1-scores, and Gradient-weighted Class Activation Mapping. Under the reported experimental protocol, the model achieved a validation accuracy of 99.74% and a macro-average F1-score of 0.8907. Bacterial spot and early blight showed comparatively strong classification performance, whereas target spot was the most difficult category to classify. Grad-CAM visualisations indicated that the model generally focused on symptom-bearing leaf regions. Comparisons with selected ResNet and VGG architectures showed favourable results for the proposed model on the same dataset. Independent external validation and deployment-oriented testing remain necessary before broader field applicability can be established.
Rice productivity is significantly affected by leaf diseases that reduce crop yield and quality. Conventional disease identification methods rely on manual observation, which is often time-consuming, subjective, and prone to misclassification due to similarities in visual symptoms. This study proposes an automated image-based classification system to detect rice leaf diseases accurately and efficiently. The system utilizes a deep learning model based on convolutional neural networks to classify rice leaf images into three disease categories: neck blast, leaf blight, and rice hispa. A dataset consisting of 3,631 images was used, with 80% allocated for training, 10% for validation, and 10% for testing. Image preprocessing techniques, including resizing, normalization, and augmentation, were applied to improve model performance and generalization. The experimental results show that the proposed model achieved a testing accuracy of 97.80%, with high precision, recall, and F1-score across all classes. The trained model was then deployed into a web-based system that enables users to upload images and obtain real-time classification results. The findings demonstrate that the proposed system provides a reliable and practical solution for early disease detection, supporting precision agriculture and improving decision-making for farmers. The system also offers potential for further development into mobile and integrated smart farming platforms.
Dian Widiarti, Olabode D. Ibini· JOKI: Journal of Computing a...· 0 citations
Farmers produce a lot of tomatoes, hence early identification and tomato leaf disease diagnosis can increase production. The information used in this research is public ally available tomato leaf image dataset with multiple disease classifications to create and improve the model. This work employs a methodical preparation workflow that includes picture scaling, normalization, and augmentation to increase resilience to variations in illumination, background clutter, and leaf texture. Two improved deep learning models, Residual and MobileNetV2, the processed pictures to enable efficient mining of significant spatial and texture information related to illness patterns. The experiment's results demonstrate the framework's utility and reliability: MobileNetV2 showed precision of 95.11, accuracy of 95.01, F1 of 95.00, and recall of 95.01 in contrast. The residual model showed 97.76 accuracy, 97.78 precision, 97.76 recall, and 97.64 F1. For every evaluation metric, these perform better than conventional DL architectures like VGG16, CNN, and ResNet. impacting early identification and improved agricultural performance.
Drashi Jain, Satendra Kumar Jain· 2026 International Conferenc...· 0 citations
Background: The necessity for effective and precise disease detection techniques is highlighted by the rising demand for legumes. Convolutional Neural Networks (CNNs), a type of deep learning, provide a potent way to diagnose plant diseases. CNNs make it possible to accurately identify illnesses in real time by quickly evaluating enormous amounts of plant pictures. By giving farmers proactive tools for monitoring crop health, cutting losses and enhancing food quality, automated detection systems can improve agricultural practices.
Methods: To categorize bean leaves, this study suggests a deep learning-based method utilizing the ResNet-20 model. To increase model generalization and lessen overfitting, data augmentation techniques such as scaling, rotation and flipping were employed. The model was trained on a dataset of labelled images and its performance was assessed using categorization metrics, confusion matrix, ROC curve and Matthews Correlation Coefficient.
Result: The ResNet-20 model’s test accuracy was 76.15%. Additional performance indicators such as metrices demonstrated the model’s reliable classification abilities. The ROC curve further illustrated the model’s exceptional ability to differentiate between healthy and unhealthy leaves.
B. D. Patil, Geetika Parmar, Manisha Shinde-Pawar et al.· Indian Journal of Agricultur...· 0 citations
Cotton leaf diseases significantly affect crop productivity, fiber quality, and agricultural sustainability, making early and accurate disease diagnosis essential for effective crop management. Manual disease identification is time-consuming, subjective, and dependent on expert knowledge, highlighting the need for automated and intelligent diagnostic systems. This study proposes a CNN-based deep learning framework for automated multiclass classification of cotton leaf diseases using digital image analysis. The proposed framework is trained on a curated dataset comprising healthy and diseased cotton leaf images representing multiple disease categories. CNN architecture performs hierarchical feature extraction, automatically learning discriminative spatial representations from input images without manual feature engineering. The network is trained for 100 epochs using optimized hyperparameters to achieve robust feature learning while ensuring high generalization capability. Model performance is quantitatively evaluated using standard classification metrics, including accuracy, precision, recall, F1-score, and confusion matrix analysis, providing a comprehensive assessment of classification effectiveness. Experimental results demonstrate that the proposed framework achieves reliable disease recognition with high classification accuracy and minimal inter-class confusion, validating the effectiveness of CNN-based feature learning for agricultural image analysis. The automated system enables rapid and accurate disease diagnosis, supporting timely intervention, optimized crop protection strategies, and improved decision-making in precision agriculture. Future work will focus on improving model robustness through the integration of larger and more diverse datasets, advanced data augmentation techniques, transfer learning, and hybrid deep learning architectures. Furthermore, deployment of the proposed framework on mobile, edge, and web-based platforms, together with IoT technologies, can facilitate real-time field monitoring and disease surveillance. Overall, the proposed CNN framework provides a scalable, computationally efficient, and intelligent solution for automated cotton leaf disease classification, contributing to the advancement of AI-driven precision agriculture and sustainable crop management.
Sonali Kamra, Vijay Laxmi· International Journal of Res...· 0 citations
Eggplant is an important horticultural crop whose productivity is often affected by various leaf diseases that reduce crop quality and yield. Manual identification of plant diseases relies heavily on human observation and experience, making it time-consuming and prone to misclassification, especially when symptoms appear visually similar. The system employed a convolutional neural network/transfer learning model to identify eggplant leaf diseases accurately and efficiently. The system utilizes a deep learning model trained on a dataset of 3,551 leaf images categorized into seven disease classes and one healthy class. Image preprocessing and augmentation techniques were applied to improve model performance and generalization. Experimental evaluation showed that the proposed model achieved a testing accuracy of approximately 82% with balanced precision and recall across all categories, indicating stable classification performance. The trained model was integrated into a web-based application that allows users to upload leaf images and obtain real-time diagnostic results along with recommended handling information. The findings demonstrate that the proposed system provides a practical and reliable solution for early disease detection and supports more efficient agricultural management. Future development may include expanding dataset diversity, improving model robustness, and deploying mobile-based applications to enhance accessibility and scalability in precision agriculture.
Abu Tholib, Moh. Ainol Yaqin· JOKI: Journal of Computing a...· 0 citations
Plant disease significantly influences agricultural production performance, particularly in valuable crops such as potatoes and tomatoes. The disease detection should be precise and automated in order to improve crop management and sustainable agriculture. Leaf disease classification for a number of diseases via multi-class approach using EfficientNetB3 architecture is proposed as the optimal structure of deep learning. Weights trained on ImageNet are used to facilitate transfer learning and feature extraction efficiency, while customized classifiers are designed to provide improved generalization, including batch normalization and dropout techniques. Input data diversity and robustness is provided via data augmentation techniques (rotation, translation, and horizontal flips). The adaptive learning approach together with Adamax optimizer are used to train the classifier. The classifier was tested using the PotatoTomato dataset from the PlantVillage repository which consists of six disease and healthy categories. The achieved accuracy score was 99.25%, while achieving precision and recall of various categories at 100%. The good performance in multi-classification proves the soundness of the classifier.
Gaurav Tuteja, Syed Nawaz Pasha, Tushar Sharma et al.· 2026 International Conferenc...· 0 citations