Tomato Leaf Disease Detection with Precaution Using EfficientNet-B3

Tomato cultivation is crucial to global agriculture, yet it is highly vulnerable to a variety of diseases and pests that significantly affect both yield and quality. In particular, fungal infections, viral pathogens like Yellow Leaf Curl Virus, and bacterial diseases such as Late Blight, are common threats that impair production. Effective and timely detection of these diseases is essential to minimize crop loss and ensure higher yields. This research explores the application of the EfficientNet-B3 deep learning model for the classification and early detection of tomato leaf diseases. By employing state-of-the-art image processing techniques and deep learning architectures, this study aims to develop a robust detection system that identifies diseased tomato leaves with high accuracy. We used the Plant Village dataset for training and testing purposes, achieving an overall classification accuracy of 99%, surpassing several traditional machine learning models. In addition to detecting and classifying diseases, the proposed system provides precautionary tips to help farmers take timely actions and prevent further spread. This scalable and computationally efficient solution leverages FastAPI to ensure high performance, assisting in early disease detection and management for improved agricultural outcomes. Furthermore, this paper discusses the importance of dataset quality, compares the model's performance with other deep learning approaches, and highlights future avenues for enhancing classification systems in agricultural settings.