Rice Leaf Disease Detection Using InceptionV3 Transfer Learning Model

This research presents a novel application of the InceptionV3 model for detecting rice leaf diseases, which holds significant importance in agricultural disease control. The present study employs transfer learning methodologies to improve the efficacy of the InceptionV3 architecture, which has undergone prior training on a diverse dataset. This study aims to precisely classify 14 distinct categories of diseases that impact the leaves of rice plants. The model's training process consists of 100 epochs, with a batch size of 40. The training procedure employs the Adam and Adamax optimization algorithms with the default learning rate. The findings of this study exhibit great promise, as the InceptionV3model, which was constructed as part of this research, demonstrated an exceptional peak accuracy of 96%. The model's capacity to effectively detect complex patterns and differentiate complex aspects in photographs significantly underscores its potential to revolutionize the identification of rice illnesses. This technical advance significantly impacts the agriculture industry by facilitating the rapid identification of diseases and the prompt application of mitigation measures. The current investigation represents a notable progression in rice disease management, intending to improve the efficiency and accuracy of instruments employed in this domain. The research discoveries possess the capacity to augment crop yield and make a significant contribution to global food security.