AlexNet Convolutional Neural Network for Disease Detection and Classification of Tomato Leaf

• DOI: 10.3390/electronics11060951
• Publication Date: 2022-03-21T01:26:22Z
• Authors: Hsing-Chung Chen; Agung Mulyo Widodo; Andika Wisnujati; Mosiur Rahaman; Jerry Chun-Wei Lin; Liukui Chen; Chien-Erh Weng
• Keywords: RGB color model

With limited retrieval of reserves and restricted capability in plant pathology, automation processes becomes essential. All over the world, farmers are struggling to prevent various harm from bacteria or pathogens such as viruses, fungi, worms, protozoa, insects. Deep learning is currently widely used across a wide range applications, including desktop, web, mobile. In this study, authors attempt implement function AlexNet modification architecture-based CNN on Android platform predict tomato diseases based leaf image. A dataset with 18,345 training data 4,585 testing was create predictive model. The information separated into ten labels for diseases, each 64 × RGB pixels. best model using Adam optimizer realizing rate 0.0005, number epochs 75, batch size 128, an uncompromising cross-entropy loss function, has high accuracy average 98%, strictness 0.98, recall value 0.99, F1-count 0.98 0.1331, so that classification results good very precise.