Fine-tuning for Transfer Learning of ResNet152 for Disease Identification in Tomato Leaves

Abstract

AbstractPlants provide a significant portion of the world’s food supply. Tomato is the most popular plant which is cultivated worldwide. The tomato leaf disease is the primary factor in productivity loss but can be avoided by monitoring regularly. Detection of tomato leaf diseases using pre-trained deep learning models can help to reduce the severity of the disease identification. However, instead of using a pre-trained model directly, there is an optional step to fine-tune the model in transfer learning, which improves the model performance. The examination of fine-tuning the model with four various scenarios of transfer learning and the art of employing pre-trained models were suggested in this work. Experiments were done using the pre-trained model ResNet152 on tomato leaf disease identification.KeywordsTransfer learningResNet152Pre-trainedDeep learningFine-tuning