Automated Detection of Tomato Leaf Disease Using Convolutional Neural Networks: Evaluation of GoogLeNet, AlexNet, and Custom CNN Architectures

Tomato leaf diseases are major global agricultural productivity barriers. It is crucial to diagnose infections early and accurately in order to successfully control them and boost agricultural productivity. This study focuses on automated disease prediction of tomato leaves using convolutional neural networks (CNNs). The study assessed three widely used CNN architectures: GoogLeNet, AlexNet, and a custom-built CNN using an extensive dataset of images of tomato leaves. The results indicate that CNN based methods produce encouraging outcomes and the GoogLeNet architecture obtains the maximum accuracy of 92%. This means deep learning methods may be used to successfully identify tomato plant diseases. This can be portrayed as the application of CNNs in solving realworld issues in agriculture.