Back

On Using Transfer Learning For Plant Disease Detection

Sagar, A.; J, D.

2020-05-23 plant biology
10.1101/2020.05.22.110957 bioRxiv
Show abstract

Deep neural networks has been highly successful in image classification problems. In this paper, we show how neural networks can be used for plant disease recognition in the context of image classification. We have used publicly available Plant Village dataset which has 38 classes of diseases. Hence, the problem that we have addressed is a multi class classification problem. We compared five different architectures including VGG16, ResNet50, InceptionV3, InceptionResNet and DenseNet169 as the backbones for our work. We found that ResNet50 achieves the best result on the test set. For evaluation, we used metrics: accuracy, precision, recall, F1 score and class wise confusion metric. Our model achieves the best of results using ResNet50 with accuracy of 0.982, precision of 0.94, recall of 0.94 and F1 score of 0.94.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.