LAD-Net: A Novel Light Weight Model for Early Apple Leaf Pests and Diseases Classification
Xianyu Zhu¹, Jinjiang Li¹, Runchang Jia¹, Bin Liu¹, Zhuohan Yao¹, Aihong Yuan², Yingqiu Huo¹, Haixi Zhang¹
¹College of Information Engineering, Northwest A&F University, Yangling, China
²College of Mechanical and Electronic Engineering, Northwest A&F University, Yangling, China
Published in IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2023
Keywords: apple leaf diseases, asymmetric convolution, real-time inference, mobile deployment
Abstract
Early detection of apple leaf diseases is crucial for agricultural production. However, existing deep learning models are often too large and computationally expensive for real-time deployment on mobile devices. This paper proposes LAD-Net, a novel lightweight model for early apple leaf pests and diseases classification. The model incorporates asymmetric convolution and dilated convolution to reduce parameters while maintaining high accuracy. Experimental results show that LAD-Net achieves 98.58% accuracy on the apple leaf disease dataset with only 1.25MB model size, making it suitable for real-time inference on mobile devices. The model also demonstrates superior performance compared to existing lightweight models in terms of accuracy, model size, and inference speed.
Recommended citation: Xianyu Zhu, Jinjiang Li, Runchang Jia, Bin Liu, Zhuohan Yao, Aihong Yuan, Yingqiu Huo, Haixi Zhang, "LAD-Net: A Novel Light Weight Model for Early Apple Leaf Pests and Diseases Classification." IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2023.
@article{zhu2023lad,
title={LAD-Net: A Novel Light Weight Model for Early Apple Leaf Pests and Diseases Classification},
author={Zhu, Xianyu and Li, Jinjiang and Jia, Runchang and Liu, Bin and Yao, Zhuohan and Yuan, Aihong and Huo, Yingqiu and Zhang, Haixi},
journal={IEEE/ACM Transactions on Computational Biology and Bioinformatics},
year={2023},
publisher={IEEE}
}