Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution
In this paper, an improved VGG16 combined with depthwise group over-parameterized convolution (DGOVGG16) model is proposed to realize automatic furniture image classification. Firstly, depthwise over-parameterized convolution combined with group convolution is combined to construct depthwise group o...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-11-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/11/23/3889 |
| _version_ | 1797463424180944896 |
|---|---|
| author | Han Ye Xiaodong Zhu Chengyang Liu Linlin Yang Aili Wang |
| author_facet | Han Ye Xiaodong Zhu Chengyang Liu Linlin Yang Aili Wang |
| author_sort | Han Ye |
| collection | DOAJ |
| description | In this paper, an improved VGG16 combined with depthwise group over-parameterized convolution (DGOVGG16) model is proposed to realize automatic furniture image classification. Firstly, depthwise over-parameterized convolution combined with group convolution is combined to construct depthwise group over-parameterized convolution, which is introduced to the VGG 16 model for reducing the number of parameters of the overall model while extracting more sufficient semantic features of furniture images. Then, this paper uses the ReLU activation function in the former part of the neural network to reduce the correlation between parameters and accelerate the weight update speed of the former part of the model. Meantime, the proposed model applies Leaky-ReLU activation function in the last layer to avoid the problem that some neurons do not update. Compared with the six furniture image classification methods based on MobileNetV2, AlexNet, ShuffleNetv2, GoogleNet, VGG 16 and GVGG16, the experimental results show the proposed DGOVGG16 with average accuracy (AA) of 95.51% has better classification performance. |
| first_indexed | 2024-03-09T17:50:30Z |
| format | Article |
| id | doaj.art-8bf03c85de2a42f48f02dab8d517b9bd |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-09T17:50:30Z |
| publishDate | 2022-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-8bf03c85de2a42f48f02dab8d517b9bd2023-11-24T10:47:14ZengMDPI AGElectronics2079-92922022-11-011123388910.3390/electronics11233889Furniture Image Classification Based on Depthwise Group Over-Parameterized ConvolutionHan Ye0Xiaodong Zhu1Chengyang Liu2Linlin Yang3Aili Wang4College of Materials Science and Engineering, Northeast Forestry University, Harbin 150040, ChinaCollege of Materials Science and Engineering, Northeast Forestry University, Harbin 150040, ChinaHeilongjiang Province Key Laboratory of Laser Spectroscopy Technology and Application, Harbin University of Science and Technology, Harbin 150080, ChinaHeilongjiang Province Key Laboratory of Laser Spectroscopy Technology and Application, Harbin University of Science and Technology, Harbin 150080, ChinaHeilongjiang Province Key Laboratory of Laser Spectroscopy Technology and Application, Harbin University of Science and Technology, Harbin 150080, ChinaIn this paper, an improved VGG16 combined with depthwise group over-parameterized convolution (DGOVGG16) model is proposed to realize automatic furniture image classification. Firstly, depthwise over-parameterized convolution combined with group convolution is combined to construct depthwise group over-parameterized convolution, which is introduced to the VGG 16 model for reducing the number of parameters of the overall model while extracting more sufficient semantic features of furniture images. Then, this paper uses the ReLU activation function in the former part of the neural network to reduce the correlation between parameters and accelerate the weight update speed of the former part of the model. Meantime, the proposed model applies Leaky-ReLU activation function in the last layer to avoid the problem that some neurons do not update. Compared with the six furniture image classification methods based on MobileNetV2, AlexNet, ShuffleNetv2, GoogleNet, VGG 16 and GVGG16, the experimental results show the proposed DGOVGG16 with average accuracy (AA) of 95.51% has better classification performance.https://www.mdpi.com/2079-9292/11/23/3889furniture image classificationdeep learningVGG16group convolutiondepthwise over-parameterized convolution |
| spellingShingle | Han Ye Xiaodong Zhu Chengyang Liu Linlin Yang Aili Wang Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution Electronics furniture image classification deep learning VGG16 group convolution depthwise over-parameterized convolution |
| title | Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution |
| title_full | Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution |
| title_fullStr | Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution |
| title_full_unstemmed | Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution |
| title_short | Furniture Image Classification Based on Depthwise Group Over-Parameterized Convolution |
| title_sort | furniture image classification based on depthwise group over parameterized convolution |
| topic | furniture image classification deep learning VGG16 group convolution depthwise over-parameterized convolution |
| url | https://www.mdpi.com/2079-9292/11/23/3889 |
| work_keys_str_mv | AT hanye furnitureimageclassificationbasedondepthwisegroupoverparameterizedconvolution AT xiaodongzhu furnitureimageclassificationbasedondepthwisegroupoverparameterizedconvolution AT chengyangliu furnitureimageclassificationbasedondepthwisegroupoverparameterizedconvolution AT linlinyang furnitureimageclassificationbasedondepthwisegroupoverparameterizedconvolution AT ailiwang furnitureimageclassificationbasedondepthwisegroupoverparameterizedconvolution |