A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism
To address the challenges of detecting a large number of objects and a high proportion of small objects in aerial drone imagery, we proposed an aerial dense small object detection algorithm called Global Normalization Attention Mechanism You Only Look Once (GNYL) based on the Global Normalization At...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/21/11760 |
| _version_ | 1827765838100299776 |
|---|---|
| author | Huixin Wu Yang Zhu Liuyi Wang |
| author_facet | Huixin Wu Yang Zhu Liuyi Wang |
| author_sort | Huixin Wu |
| collection | DOAJ |
| description | To address the challenges of detecting a large number of objects and a high proportion of small objects in aerial drone imagery, we proposed an aerial dense small object detection algorithm called Global Normalization Attention Mechanism You Only Look Once (GNYL) based on the Global Normalization Attention Mechanism. In the backbone network of GNYL, we embedded a GNAM (Global Normalization Attention Mechanism) that explores channel attention features and spatial attention features from input features in a concatenated manner. It utilizes batch normalization’s scale factors to suppress irrelevant channels or pixels. Furthermore, the spatial attention sub-module introduces a three-dimensional arrangement with a multi-layer perceptron to reduce information loss and amplify global interaction representation. Finally, the computed attention weights are weighted to form the global normalized attention weights, which increases the utilization of effective information in input feature channels and spatial dimensions. We have optimized the backbone network, feature enhancement network, and detection heads to improve detection accuracy while ensuring a lightweight detection network. Specifically, we have added a small object detection layer to enhance the localization accuracy for the abundant small objects in aerial imagery. The algorithm’s performance was evaluated using the publicly available VisDrone2019 dataset. Compared to the baseline network YOLOv8l, GNYL achieved a 7.2% improvement in mAP<sub>0.5</sub> and a 5.0% improvement in mAP<sub>0.95</sub>. Compared to CDNet, GNYL showed a 14.5% improvement in mAP<sub>0.5</sub> and a 9.1% improvement in mAP<sub>0.95</sub>. These experimental results demonstrate the strong practicality of the GNYL object detection network for detecting dense small objects in the aerial imagery captured by unmanned aerial vehicles. |
| first_indexed | 2024-03-11T11:33:30Z |
| format | Article |
| id | doaj.art-4a83ca48bc334ca6b06b80687798816c |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T11:33:30Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-4a83ca48bc334ca6b06b80687798816c2023-11-10T14:58:37ZengMDPI AGApplied Sciences2076-34172023-10-0113211176010.3390/app132111760A Dense Small Object Detection Algorithm Based on a Global Normalization Attention MechanismHuixin Wu0Yang Zhu1Liuyi Wang2School of Information Engineering, North China University of Water Resources and Electric Power, No. 136 Jinshui East Road, Zhengzhou 450046, ChinaSchool of Information Engineering, North China University of Water Resources and Electric Power, No. 136 Jinshui East Road, Zhengzhou 450046, ChinaSchool of Information Engineering, North China University of Water Resources and Electric Power, No. 136 Jinshui East Road, Zhengzhou 450046, ChinaTo address the challenges of detecting a large number of objects and a high proportion of small objects in aerial drone imagery, we proposed an aerial dense small object detection algorithm called Global Normalization Attention Mechanism You Only Look Once (GNYL) based on the Global Normalization Attention Mechanism. In the backbone network of GNYL, we embedded a GNAM (Global Normalization Attention Mechanism) that explores channel attention features and spatial attention features from input features in a concatenated manner. It utilizes batch normalization’s scale factors to suppress irrelevant channels or pixels. Furthermore, the spatial attention sub-module introduces a three-dimensional arrangement with a multi-layer perceptron to reduce information loss and amplify global interaction representation. Finally, the computed attention weights are weighted to form the global normalized attention weights, which increases the utilization of effective information in input feature channels and spatial dimensions. We have optimized the backbone network, feature enhancement network, and detection heads to improve detection accuracy while ensuring a lightweight detection network. Specifically, we have added a small object detection layer to enhance the localization accuracy for the abundant small objects in aerial imagery. The algorithm’s performance was evaluated using the publicly available VisDrone2019 dataset. Compared to the baseline network YOLOv8l, GNYL achieved a 7.2% improvement in mAP<sub>0.5</sub> and a 5.0% improvement in mAP<sub>0.95</sub>. Compared to CDNet, GNYL showed a 14.5% improvement in mAP<sub>0.5</sub> and a 9.1% improvement in mAP<sub>0.95</sub>. These experimental results demonstrate the strong practicality of the GNYL object detection network for detecting dense small objects in the aerial imagery captured by unmanned aerial vehicles.https://www.mdpi.com/2076-3417/13/21/11760computer visionsmall object detectionYOLOv8attention mechanismunmanned aerial vehicle |
| spellingShingle | Huixin Wu Yang Zhu Liuyi Wang A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism Applied Sciences computer vision small object detection YOLOv8 attention mechanism unmanned aerial vehicle |
| title | A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism |
| title_full | A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism |
| title_fullStr | A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism |
| title_full_unstemmed | A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism |
| title_short | A Dense Small Object Detection Algorithm Based on a Global Normalization Attention Mechanism |
| title_sort | dense small object detection algorithm based on a global normalization attention mechanism |
| topic | computer vision small object detection YOLOv8 attention mechanism unmanned aerial vehicle |
| url | https://www.mdpi.com/2076-3417/13/21/11760 |
| work_keys_str_mv | AT huixinwu adensesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism AT yangzhu adensesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism AT liuyiwang adensesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism AT huixinwu densesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism AT yangzhu densesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism AT liuyiwang densesmallobjectdetectionalgorithmbasedonaglobalnormalizationattentionmechanism |