Convolutional Neural Networks for Foliar Disease Diagnosis in Horticulture Crops: A Critical Review

Abstract

The research aims to examine further how to apply advanced machine learning models and deep learning models, especially convolutional neural network models (CNN), in detecting disease in both tomato and potato plants in advance. It emphasizes how various factors, such as disease-producing micro-organisms and environmental components in addition to less varying datasets, especially when using datasets such as PlantVillage, which have shown vast usage in various models in various research worldwide, can particularly influence the accuracy of disease detection models in crops such as potatoes and tomatoes using CNN models talked of in this study. CNN models, especially AlexNet models, ResNet models, and EfficientNet models, which show high accuracy in classifying different kinds of disease in various kinds of crops using 94–99.75% accuracy levels, are discussed in this study as a basis of further research to come up with more efficient models in promoting both machine learning and various uses of CNN models in disease management in promoting food security in many countries worldwide.