A comparative study of fine-tuning deep learning models for plant disease identification

• Published in: Computers and electronics in agriculture 2019-06, Vol.161, p.272-279
• Main Authors: Too, Edna Chebet, Yujian, Li, Njuki, Sam, Yingchun, Liu
• Format: Article
• Language: English
• Subjects: Accuracy; Artificial neural networks; Comparative studies; Computing time; Convolutional neural networks; Deep learning; Fine-tuning; Image classification; Image recognition; Machine learning; Performance degradation; Plant disease identification; Plant diseases; Tuning
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2018.03.032

•A comparative study of state-of-the-art deep learning for plants disease detection using images of leaves.•The results show that deeper models are not only accurate but have fewer number of parameters.•DenseNet model perform better than other models studied with no signs of overfitting and performance deterioration.•DenseNet achieved an accuracy of 99.75%. Deep learning has recently attracted a lot of attention with the aim to develop a quick, automatic and accurate system for image identification and classification. In this work, the focus was on fine-tuning and evaluation of state-of-the-art deep convolutional neural network for image-based plant disease classification. An empirical comparison of the deep learning architecture is done. The architectures evaluated include VGG 16, Inception V4, ResNet with 50, 101 and 152 layers and DenseNets with 121 layers. The data used for the experiment is 38 different classes including diseased and healthy images of leafs of 14 plants from plantVillage. Fast and accurate models for plant disease identification are desired so that accurate measures can be applied early. Thus, alleviating the problem of food security. In our experiment, DenseNets has tendency’s to consistently improve in accuracy with growing number of epochs, with no signs of overfitting and performance deterioration. Moreover, DenseNets requires a considerably less number of parameters and reasonable computing time to achieve state-of-the-art performances. It achieves a testing accuracy score of 99.75% to beat the rest of the architectures. Keras with Theano backend was used to perform the training of the architectures.