Invading pests and diseases always degrade the quality and quantity of plants. Early and accurate identification of plant diseases is critical for plant health and growth. This work proposes a smartphone-based solution using a Vision Transformer (ViT) model for identifying healthy plants and unhealthy plants with diseases. The collected dataset of tomato leaves was used to collectively train Vision Transformer and Inception V3-based deep learning (DL) models to differentiate healthy and diseased plants. These models detected 10 different tomato disease classes from the dataset containing 10,010 images. The performance of the two DL models was compared. This work also presents a smartphone-based application (Android App) using a ViT-based model, which works on the basis of the self-attention mechanism and yielded a better performance (90.99% testing) than Inception V3 in our experimentation. The proposed ViT-SmartAgri is promising and can be implemented on a colossal scale for smart agriculture, thus inspiring future work in this area.

Shikhar Kr. Sarma, Ratul Deka, Bhatima Baro, Vaskar Deka, Umesh Deka, Mirzanur Rahman, Sarmah Satyajit, Kuwali Talukdar, Kishore Kashyap. Proceedings of the 13th Global Wordnet Conference. 2025.

With the proliferation of digital content, efficient image matching in natural image databases has become paramount. Traditional image matching techniques, while effective to a certain extent, face challenges in dealing with the high variability inherent in natural images. This research delves into the application of deep learning models, particularly Convolutional Neural Networks (CNNs), Siamese Networks, and Triplet Networks, to address these challenges. We introduce various techniques to enhance efficiency, such as data augmentation, transfer learning, dimensionality reduction, efficient sampling, and the amalgamation of traditional computer vision strategies with deep learning. Our experimental results, garnered from specific dataset, demonstrate significant improvements in image matching efficiency, as quantified by metrics like precision, recall, F1-Score, and matching time. The findings underscore the potential of deep learning as a transformative tool for natural image database matching, setting the stage for further research and optimization in this domain.