Semi-Supervised Noisy Student Pre-training on EfficientNet Architectures for Plant Pathology Classification
This work provides an incremental improvement in plant disease identification for agricultural applications.
This paper addresses plant pathology classification from leaf images. The authors achieved a test score of 0.982 using an ensembled Noisy Student model on EfficientNet architectures, improving upon a 0.962 score from EfficientNet alone.
In recent years, deep learning has vastly improved the identification and diagnosis of various diseases in plants. In this report, we investigate the problem of pathology classification using images of a single leaf. We explore the use of standard benchmark models such as VGG16, ResNet101, and DenseNet 161 to achieve a 0.945 score on the task. Furthermore, we explore the use of the newer EfficientNet model, improving the accuracy to 0.962. Finally, we introduce the state-of-the-art idea of semi-supervised Noisy Student training to the EfficientNet, resulting in significant improvements in both accuracy and convergence rate. The final ensembled Noisy Student model performs very well on the task, achieving a test score of 0.982.