TomatoCare: A Mobile Deep Learning System for Leaf Disease Detection

Abstract

Tomato crops are highly vulnerable to leaf diseases, which can cause major yield losses if not detected early. In this work, we present a mobile application that helps farmers identify tomato leaf diseases quickly and reliably. The system uses a three-stage deep learning pipeline: YOLOv8 for detecting leaves in images, a lightweight CNN model for checking if leaves are healthy or diseased, and a multi-class CNN for identifying the exact disease. A simple question-based verification step further improves trust in predictions. Designed for offline use, the app delivers accurate, fast, and farmer-friendly disease diagnosis directly on smartphones. © 2026 IEEE.