A Custom CNN Model for Tomato Leaf Disease Detection

Tomato cultivation is playing a vital role both in universal agriculture and economy. However, the presence of various types of diseases poses a significant threat to the tomato plants and result in substantial reduction in crop yield. Therefore, detecting and managing these diseases at an early age is very crucial in order to minimize damages. Due to technological advancements, Convolutional Neural Networks have become known as an effective tool for recognition of plant diseases because of their remarkable capability of recognizing and classifying of images. This research work proposed the use of a Convolutional Neural Network(CNN) model that is built from scratch. The proposed model is made up of 14 layers and showed promising result with high accuracy of 97.91%, a precision of 96.51%, with 97.47% and 96.91% for recall and F1-score respectively. The proposed custom model also outperforms a transfer learning model and other deep learning models which indicates how effective and robust the model is when it comes to tomato disease detection.