LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices
ShuffleNetV2 is a prominent player in the field of lightweight networks and has significant implications for the development of lightweight networks and edge computing. However, it has limitations, as its accuracy falls short compared to other larger models, and it is not friendly to datasets. It lo...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10286272/ |
| _version_ | 1797648680171339776 |
|---|---|
| author | Hui Zhang Xiaoyang Zhu Bo Li Zhen Guan Weimin Che |
| author_facet | Hui Zhang Xiaoyang Zhu Bo Li Zhen Guan Weimin Che |
| author_sort | Hui Zhang |
| collection | DOAJ |
| description | ShuffleNetV2 is a prominent player in the field of lightweight networks and has significant implications for the development of lightweight networks and edge computing. However, it has limitations, as its accuracy falls short compared to other larger models, and it is not friendly to datasets. It loses its advantage for small-sized images with fewer channels. In this study, we analysed the design structure of ShuffleNetV2 and found room for improvement in its computational complexity and accuracy. To further improve its performance, we first upgraded the ShuffleNetV2 network structure based on the lightweight network design criteria and constructed a new network model. Second, we introduced a novel attention module named the Adaptive Pooling Attention Module (APAM) and integrated it with the new network model, constructing a high-performance model referred to as LA-ShuffleNet. Then, we proposed a convolution operation acceleration strategy called Pack. Finally, we combined the two and conducted corresponding tests on the Windows and JETSON platforms. Extensive experiments indicate that our proposed model not only exhibits substantial improvements over the baseline model but also achieves noteworthy enhancements on the ImageNet dataset, with a rise of 1.4% in Top-1 accuracy and 3.6% in Top-5 accuracy, coupled with a reduction of 0.7M parameters. Moreover, its performance surpasses that of certain prevalent lightweight networks, such as MobileNet. |
| first_indexed | 2024-03-11T15:34:27Z |
| format | Article |
| id | doaj.art-c5695c4956884258a445977ed9880d21 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-11T15:34:27Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-c5695c4956884258a445977ed9880d212023-10-26T23:01:19ZengIEEEIEEE Access2169-35362023-01-011111668411669410.1109/ACCESS.2023.332471310286272LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing DevicesHui Zhang0https://orcid.org/0000-0001-8467-8852Xiaoyang Zhu1https://orcid.org/0009-0006-1997-9353Bo Li2Zhen Guan3Weimin Che4https://orcid.org/0009-0002-3136-5486School of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang, ChinaSchool of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang, ChinaSchool of Industrial Software, Henan University of Engineering, Zhengzhou, ChinaSchool of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang, ChinaSchool of Mechanical Engineering, Jiangsu University of Science and Technology, Zhenjiang, ChinaShuffleNetV2 is a prominent player in the field of lightweight networks and has significant implications for the development of lightweight networks and edge computing. However, it has limitations, as its accuracy falls short compared to other larger models, and it is not friendly to datasets. It loses its advantage for small-sized images with fewer channels. In this study, we analysed the design structure of ShuffleNetV2 and found room for improvement in its computational complexity and accuracy. To further improve its performance, we first upgraded the ShuffleNetV2 network structure based on the lightweight network design criteria and constructed a new network model. Second, we introduced a novel attention module named the Adaptive Pooling Attention Module (APAM) and integrated it with the new network model, constructing a high-performance model referred to as LA-ShuffleNet. Then, we proposed a convolution operation acceleration strategy called Pack. Finally, we combined the two and conducted corresponding tests on the Windows and JETSON platforms. Extensive experiments indicate that our proposed model not only exhibits substantial improvements over the baseline model but also achieves noteworthy enhancements on the ImageNet dataset, with a rise of 1.4% in Top-1 accuracy and 3.6% in Top-5 accuracy, coupled with a reduction of 0.7M parameters. Moreover, its performance surpasses that of certain prevalent lightweight networks, such as MobileNet.https://ieeexplore.ieee.org/document/10286272/Lightweight networkattention moduleedge computingShuffleNet |
| spellingShingle | Hui Zhang Xiaoyang Zhu Bo Li Zhen Guan Weimin Che LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices IEEE Access Lightweight network attention module edge computing ShuffleNet |
| title | LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices |
| title_full | LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices |
| title_fullStr | LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices |
| title_full_unstemmed | LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices |
| title_short | LA-ShuffleNet: A Strong Convolutional Neural Network for Edge Computing Devices |
| title_sort | la shufflenet a strong convolutional neural network for edge computing devices |
| topic | Lightweight network attention module edge computing ShuffleNet |
| url | https://ieeexplore.ieee.org/document/10286272/ |
| work_keys_str_mv | AT huizhang lashufflenetastrongconvolutionalneuralnetworkforedgecomputingdevices AT xiaoyangzhu lashufflenetastrongconvolutionalneuralnetworkforedgecomputingdevices AT boli lashufflenetastrongconvolutionalneuralnetworkforedgecomputingdevices AT zhenguan lashufflenetastrongconvolutionalneuralnetworkforedgecomputingdevices AT weiminche lashufflenetastrongconvolutionalneuralnetworkforedgecomputingdevices |