Lightweight CNN model for apple leaf disease identification

Xiuman Liang; Shaopin Gao; Zhendong Liu

doi:10.1117/12.2689350

31 July 2023 Lightweight CNN model for apple leaf disease identification

Xiuman Liang, Shaopin Gao, Zhendong Liu

Proceedings Volume 12747, Third International Conference on Optics and Image Processing (ICOIP 2023); 127471U (2023) https://doi.org/10.1117/12.2689350
Event: Third International Conference on Optics and Image Processing (ICOIP 2023), 2023, Hangzhou, China

Abstract

Although the increase of network depth and width can increase the accuracy of recognition, the number of parameters and computation are often large, which is not suitable for mobile device applications. To solve this problem, two lightweight CNN models are constructed to improve the feature extraction and feature reuse capability of the network by improving the Fire module of SqueezeNet network, adding a spatial attention mechanism and adding a dense connection module in the deeper layer of the network. The improved models achieved 89.60% and 94.37% recognition accuracy by training on the constructed apple disease leaf dataset, which is 2.29% and 7.75% higher than the original network, while the number of parameters of the network is only 0.9M and 2.5M. The experimental results show that the improved network achieves higher recognition accuracy while keeping the lightweight model.

Citation Download Citation

Xiuman Liang, Shaopin Gao, and Zhendong Liu "Lightweight CNN model for apple leaf disease identification", Proc. SPIE 12747, Third International Conference on Optics and Image Processing (ICOIP 2023), 127471U (31 July 2023); https://doi.org/10.1117/12.2689350

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available