Description of Invention
Proper identification of plant diseases is essential for ensuring the production of high-quality food. Plant disease detection is crucial in the study of agricultural environmental science [1]. According to some statistics, many diseases have decreased rice field production in recent years. These diseases, which are caused by bacteria, fungi, and viruses, can cost the farmers to lose a huge amount of the crops. In addition, there are some common rice diseases such as brownSpot, Hispa, and LeafBlast [2, 3]. Computer vision has become popular for detection and categorization. These approaches can diagnose diseases early before symptoms appear. Convolutional Neuron Network (CNN) is a popular deep learning architecture due to its high model capacity and ability to handle complex data [4, 5, 6]. Transfer learning in deep learning employs CNNs trained for one job to build models for other tasks. Before being retrained for target dataset tasks, models should be pretrained on a large dataset like ImageNet. The plant disease classification models such as VGGNet, ResNet, Inception V4, DenseNets, and SqueezeNet use transfer learning extensively [7, 8].