James Vincent Bacus, Caitlin Mariel Lindsay, Christian Maderazo, Gerard Ompad
Crop diseases reduce yield, affecting a nation's agricultural sector. This is more pronounced in nations such as the Philippines, where agriculture is the nation's foundation. To decrease the cost and time needed for disease detection, this study utilized Transfer Learning on three Convolutional Neural Networks (DenseNet201, EfficientNetV2M, and InceptionResNetV2) to identify three corn diseases: Common Rust, Blight, and Gray Leaf Spot. Data Augmentation was used to diversify the image dataset, resulting in 4,000 images per category. This research presents a comparison between the different architectures. It also illustrates the effect of the Soft Voting Ensemble Method. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.
Department of Computer and Information Science, University of San Carlos, Cebu, Philippines; Faculty of Department of Computer and Information Science, University of San Carlos, Cebu, Philippines
