Garbage classification algorithm based on improved EfficientNetV2 network

Jiali Cui; Han Li; Zhenghui Hu

doi:10.1117/12.2680151

31 May 2023 Garbage classification algorithm based on improved EfficientNetV2 network

Jiali Cui, Han Li, Zhenghui Hu

Proceedings Volume 12704, Eighth International Symposium on Advances in Electrical, Electronics, and Computer Engineering (ISAEECE 2023); 1270436 (2023) https://doi.org/10.1117/12.2680151
Event: 8th International Symposium on Advances in Electrical, Electronics and Computer Engineering (ISAEECE 2023), 2023, Hangzhou, China

Abstract

The background information of garbage images is complex, and the difference within the class is large. It is difficult for traditional methods to extract significant classification features from garbage images, resulting in poor classification results. Aiming at the above problems, a garbage image classification method based on attention mechanism and SPP model is proposed. First, EfficientNetV2 is used as the basic network, and an improved SE attention module is introduced to make the network pay more attention to the important features in the feature map, suppress the background information, and reduce the impact of intra class differences; Secondly, in order to deepen the network depth and integrate multi-scale information, a feature pyramid structure is introduced; Finally, the experimental results show that on the extended Huawei dataset, the classification accuracy is improved by about 4.8%, reaching 93.5%, which proves that this method has certain advantages in garbage classification tasks.

Citation Download Citation

Jiali Cui, Han Li, and Zhenghui Hu "Garbage classification algorithm based on improved EfficientNetV2 network", Proc. SPIE 12704, Eighth International Symposium on Advances in Electrical, Electronics, and Computer Engineering (ISAEECE 2023), 1270436 (31 May 2023); https://doi.org/10.1117/12.2680151

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