A Novel Hybrid Deep Learning Model for Crop Disease Detection Using BEGAN

Crop diseases are a considerable threat in the agricultural sector as they adversely affect the production and quality of agricultural products, resulting in heavy economic losses for both farmers and the country. Therefore, early identification and diagnosis of crop diseases at each stage of their lifespan is critical to protect and maximize crop yields. In this paper, we have proposed a novel deep learning model that utilizes the began to generate synthetic images of crop leaves in order to improve the network generalizability. Thereafter, a hybrid InceptionV3 + RF model is trained on real and synthetic images using transfer learning to classify crop leaves images in ten categories.