Classification of strawberry diseases and pests by improved AlexNet deep learning networks

Authors:

Abstract
To improve the classification accuracy of strawberry diseases and pests, this paper proposed an improved operator-based convolutional neural network (CNN) approach for classification of images of strawberry diseases and pests. Firstly, by using the deep learning framework of Pytorch, we fine-tuned the AlexNet model so that it was used to train the image dataset of strawberry diseases and pests. Next, combining inner product with l 2 -norm, we proposed a new operator to replace the inner product operator between input values and weights in the fully connected layers of the AlexNet model. Then the proposed operator was applied to classification of strawberry diseases and pests. By experimental verification, the proposed method on the independent test set for the classification accuracy has been considerably increased. Our source code is available at https://gitee.com/dc2019/improved-alexnet.
Funding Information
  • National Natural Science Foundation of China

This publication has 9 references indexed in Scilit: