Object Detection Network Based on Module Stack and Attention Mechanism
Currently, visual computer applications based on convolutional neural networks are rapidly developing. However, several problems remain: (1) high-quality graphics processing equipment is needed, and (2) the trained network model has several unnecessary convolution operations. These problems result i...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-08-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/12/17/3542 |
| _version_ | 1797582683744763904 |
|---|---|
| author | Xinke Dou Ting Wang Shiliang Shao Xianqing Cao |
| author_facet | Xinke Dou Ting Wang Shiliang Shao Xianqing Cao |
| author_sort | Xinke Dou |
| collection | DOAJ |
| description | Currently, visual computer applications based on convolutional neural networks are rapidly developing. However, several problems remain: (1) high-quality graphics processing equipment is needed, and (2) the trained network model has several unnecessary convolution operations. These problems result in a single-stage target detection network that often requires unnecessary computing power and is difficult to apply to equipment with insufficient computing resources. To solve these problems, based on YOLOv5, a YOLOv5-L (YOLOv5 Lightweight) network structure is proposed. This network is improved using YOLOv5. First, to enhance the inference speed of the detector on the CPU, the PP-LCNet (PaddlePaddle-Lightweight CPU Net) is employed as the backbone network. Second, the focus module is removed, and the end convolution module in the head network is replaced by a deep separable convolution module, which eliminates redundant operations and reduces the amount of computation. The experimental results show that YOLOv5-L enables a 48% reduction in model parameters and computation compared to YOLOv5, a 35% increase in operation speed, and a less than 2% reduction in accuracy, which is significant in the environment of low-performance computing equipment. |
| first_indexed | 2024-03-10T23:25:59Z |
| format | Article |
| id | doaj.art-a272371634334f8ebd49694c400399ac |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T23:25:59Z |
| publishDate | 2023-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-a272371634334f8ebd49694c400399ac2023-11-19T08:00:40ZengMDPI AGElectronics2079-92922023-08-011217354210.3390/electronics12173542Object Detection Network Based on Module Stack and Attention MechanismXinke Dou0Ting Wang1Shiliang Shao2Xianqing Cao3School of Information Engineering, Shenyang University of Chemical Technology, No. 11, 11th St., Tiexi District, Shenyang 110142, ChinaState Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, No. 114, Nanta St., Shenhe District, Shenyang 110016, ChinaState Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, No. 114, Nanta St., Shenhe District, Shenyang 110016, ChinaSchool of Information Engineering, Shenyang University of Chemical Technology, No. 11, 11th St., Tiexi District, Shenyang 110142, ChinaCurrently, visual computer applications based on convolutional neural networks are rapidly developing. However, several problems remain: (1) high-quality graphics processing equipment is needed, and (2) the trained network model has several unnecessary convolution operations. These problems result in a single-stage target detection network that often requires unnecessary computing power and is difficult to apply to equipment with insufficient computing resources. To solve these problems, based on YOLOv5, a YOLOv5-L (YOLOv5 Lightweight) network structure is proposed. This network is improved using YOLOv5. First, to enhance the inference speed of the detector on the CPU, the PP-LCNet (PaddlePaddle-Lightweight CPU Net) is employed as the backbone network. Second, the focus module is removed, and the end convolution module in the head network is replaced by a deep separable convolution module, which eliminates redundant operations and reduces the amount of computation. The experimental results show that YOLOv5-L enables a 48% reduction in model parameters and computation compared to YOLOv5, a 35% increase in operation speed, and a less than 2% reduction in accuracy, which is significant in the environment of low-performance computing equipment.https://www.mdpi.com/2079-9292/12/17/3542lightweightYOLOv5PP-LCNetobject detectionattention mechanism |
| spellingShingle | Xinke Dou Ting Wang Shiliang Shao Xianqing Cao Object Detection Network Based on Module Stack and Attention Mechanism Electronics lightweight YOLOv5 PP-LCNet object detection attention mechanism |
| title | Object Detection Network Based on Module Stack and Attention Mechanism |
| title_full | Object Detection Network Based on Module Stack and Attention Mechanism |
| title_fullStr | Object Detection Network Based on Module Stack and Attention Mechanism |
| title_full_unstemmed | Object Detection Network Based on Module Stack and Attention Mechanism |
| title_short | Object Detection Network Based on Module Stack and Attention Mechanism |
| title_sort | object detection network based on module stack and attention mechanism |
| topic | lightweight YOLOv5 PP-LCNet object detection attention mechanism |
| url | https://www.mdpi.com/2079-9292/12/17/3542 |
| work_keys_str_mv | AT xinkedou objectdetectionnetworkbasedonmodulestackandattentionmechanism AT tingwang objectdetectionnetworkbasedonmodulestackandattentionmechanism AT shiliangshao objectdetectionnetworkbasedonmodulestackandattentionmechanism AT xianqingcao objectdetectionnetworkbasedonmodulestackandattentionmechanism |