Deep Learning-Based Networks to Detect Leaf Disease in Maize and Corn

Abstract

Corn or maize possesses an exceptionally high level of genetic diversity as well as a very high capacity for productivity. It is classified with the other grains as a food group and is onto important crop grown for both people and animals. Corn, despite its crucial role, is particularly vulnerable to both pests and diseases due to its susceptibility to both. Both productionand quality are experiencing a decline as a result. The plant's tissue becomes infected, which causes the plant's growth to become stunted and its colour to change from green to yellow. The consequences of such issues are not just emotionally and physically taxing but also financially costly. Taking that part into consideration, we have worked with deep neural networks and performed experiments using ResNet50, VGG-16, EfficientNetB0, and EfficientNet B3. The best results were obtained on ResNet50 with an accuracy of 97.02 percent and a loss of 8.56 percent, as the model has been fine-tuned by utilizing a gridsearch algorithm, which searches for the best parameters to be utilized in order to achieve best possible train and test accuracy.