LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition
In recent years, the transformer model has been widely used in computer-vision tasks and has achieved impressive results. Unfortunately, these transformer-based models have the common drawback of having many parameters and a large memory footprint, causing them to be difficult to deploy on mobiles a...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/12/7/1693 |
| _version_ | 1797608096812498944 |
|---|---|
| author | Shiyong Geng Zongnan Zhu Zhida Wang Yongping Dan Hengyi Li |
| author_facet | Shiyong Geng Zongnan Zhu Zhida Wang Yongping Dan Hengyi Li |
| author_sort | Shiyong Geng |
| collection | DOAJ |
| description | In recent years, the transformer model has been widely used in computer-vision tasks and has achieved impressive results. Unfortunately, these transformer-based models have the common drawback of having many parameters and a large memory footprint, causing them to be difficult to deploy on mobiles as lightweight convolutional neural networks. To address these issues, a Vision Transformer (ViT) model, named the lightweight Vision Transformer (LW-ViT) model, is proposed to reduce the complexity of the transformer-based model. The model is applied to offline handwritten Chinese character recognition. The design of the LW-ViT model is inspired by MobileViT. The lightweight ViT model reduces the number of parameters and FLOPs by reducing the number of transformer blocks and the MV2 layer based on the overall framework of the MobileViT model. The number of parameters and FLOPs for the LW-ViT model was 0.48 million and 0.22 G, respectively, and it ultimately achieved a high recognition accuracy of 95.8% on the dataset. Furthermore, compared to the MobileViT model, the number of parameters was reduced by 53.8%, and the FLOPs were reduced by 18.5%. The experimental results show that the LW-ViT model has a low number of parameters, proving the correctness and feasibility of the proposed model. |
| first_indexed | 2024-03-11T05:38:51Z |
| format | Article |
| id | doaj.art-7303cf5fe63946be85106e2b033d7647 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-11T05:38:51Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-7303cf5fe63946be85106e2b033d76472023-11-17T16:34:17ZengMDPI AGElectronics2079-92922023-04-01127169310.3390/electronics12071693LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character RecognitionShiyong Geng0Zongnan Zhu1Zhida Wang2Yongping Dan3Hengyi Li4School of Electronic and Information, Zhongyuan University of Technology, Zhengzhou 451191, ChinaSchool of Electronic and Information, Zhongyuan University of Technology, Zhengzhou 451191, ChinaSchool of Electronic and Information, Zhongyuan University of Technology, Zhengzhou 451191, ChinaSchool of Electronic and Information, Zhongyuan University of Technology, Zhengzhou 451191, ChinaGraduate School of Science and Engineering, Ritsumeikan University, 1-1-1 Noji-higashi, Kusatsu 525-8577, JapanIn recent years, the transformer model has been widely used in computer-vision tasks and has achieved impressive results. Unfortunately, these transformer-based models have the common drawback of having many parameters and a large memory footprint, causing them to be difficult to deploy on mobiles as lightweight convolutional neural networks. To address these issues, a Vision Transformer (ViT) model, named the lightweight Vision Transformer (LW-ViT) model, is proposed to reduce the complexity of the transformer-based model. The model is applied to offline handwritten Chinese character recognition. The design of the LW-ViT model is inspired by MobileViT. The lightweight ViT model reduces the number of parameters and FLOPs by reducing the number of transformer blocks and the MV2 layer based on the overall framework of the MobileViT model. The number of parameters and FLOPs for the LW-ViT model was 0.48 million and 0.22 G, respectively, and it ultimately achieved a high recognition accuracy of 95.8% on the dataset. Furthermore, compared to the MobileViT model, the number of parameters was reduced by 53.8%, and the FLOPs were reduced by 18.5%. The experimental results show that the LW-ViT model has a low number of parameters, proving the correctness and feasibility of the proposed model.https://www.mdpi.com/2079-9292/12/7/1693transformer-based modelslightweight vision transformer (LW-ViT)offline handwritten Chinese character recognitionMV2 layer |
| spellingShingle | Shiyong Geng Zongnan Zhu Zhida Wang Yongping Dan Hengyi Li LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition Electronics transformer-based models lightweight vision transformer (LW-ViT) offline handwritten Chinese character recognition MV2 layer |
| title | LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition |
| title_full | LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition |
| title_fullStr | LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition |
| title_full_unstemmed | LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition |
| title_short | LW-ViT: The Lightweight Vision Transformer Model Applied in Offline Handwritten Chinese Character Recognition |
| title_sort | lw vit the lightweight vision transformer model applied in offline handwritten chinese character recognition |
| topic | transformer-based models lightweight vision transformer (LW-ViT) offline handwritten Chinese character recognition MV2 layer |
| url | https://www.mdpi.com/2079-9292/12/7/1693 |
| work_keys_str_mv | AT shiyonggeng lwvitthelightweightvisiontransformermodelappliedinofflinehandwrittenchinesecharacterrecognition AT zongnanzhu lwvitthelightweightvisiontransformermodelappliedinofflinehandwrittenchinesecharacterrecognition AT zhidawang lwvitthelightweightvisiontransformermodelappliedinofflinehandwrittenchinesecharacterrecognition AT yongpingdan lwvitthelightweightvisiontransformermodelappliedinofflinehandwrittenchinesecharacterrecognition AT hengyili lwvitthelightweightvisiontransformermodelappliedinofflinehandwrittenchinesecharacterrecognition |