SSRDet: Small Object Detection Based on Feature Pyramid Network

被引:5
作者
Zhang, Lijuan [1 ,2 ]
Wang, Minhui [2 ]
Jiang, Yutong [3 ]
Li, Dongming [1 ]
Zhou, Yue [3 ]
机构
[1] Wuxi Univ, Coll Internet Things Engn, Wuxi 214105, Jiangsu, Peoples R China
[2] Changchun Univ Technol, Sch Comp Sci & Engn, Changchun 130012, Jilin, Peoples R China
[3] China North Vehicle Res Inst, Beijing 100072, Peoples R China
基金
中国国家自然科学基金;
关键词
Object detection; Feature extraction; Detectors; Training; Semantics; Data augmentation; Gaussian distribution; Labeling; Small object detection; attention module; feature pyramid network; label assignment;
D O I
10.1109/ACCESS.2023.3306242
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Due to the increasing presence of small objects in videos or images from practical applications, small object identification is currently an extremely popular topic in the field of machine vision. Additionally, small object detection is still a difficult process because to small objects' issues with fuzzy appearance, limited information, occlusion, and noise present. Most existing methods mainly use feature pyramid networks to enrich shallow features using contextual features. However, due to the inconsistency of gradients between different layers of the feature pyramid network, the shallow features cannot be fully utilized resulting in the slow improvement of small object detection accuracy. To effectively improve the small object detection algorithm, we propose a new feature pyramid network-based small object detection algorithm, SSRDet. To effectively assign positive and negative sample labels and address the issue of sample scale imbalance, we first present RFLA. Then, to overcome the gradient inconsistency between various layers and enable the full utilization of the shallow features, we extend the feature pyramid network by including a scale enhancement module (SEM) and a scale selection module (SSM). Finally, we introduced the attention module (SPAM) to filter out the background noise in the shallow feature extraction to better extract small object features. We validated our method on VisDrone2019 and AI-TOD, and our method outperformed the state-of-the-art detectors.
引用
收藏
页码:96743 / 96752
页数:10
相关论文
共 36 条
[1]   SOD-MTGAN: Small Object Detection via Multi-Task Generative Adversarial Network [J].
Bai, Yancheng ;
Zhang, Yongqiang ;
Ding, Mingli ;
Ghanem, Bernard .
COMPUTER VISION - ECCV 2018, PT XIII, 2018, 11217 :210-226
[2]  
Bochkovskiy A, 2020, Arxiv, DOI arXiv:2004.10934
[3]   Cascade R-CNN: Delving into High Quality Object Detection [J].
Cai, Zhaowei ;
Vasconcelos, Nuno .
2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, :6154-6162
[4]   RRNet: A Hybrid Detector for Object Detection in Drone-captured Images [J].
Chen, Changrui ;
Zhang, Yu ;
Lv, Qingxuan ;
Wei, Shuo ;
Wang, Xiaorui ;
Sun, Xin ;
Dong, Junyu .
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS (ICCVW), 2019, :100-108
[5]  
Chen LC, 2016, Arxiv, DOI arXiv:1412.7062
[6]  
Fu C.-Y., 2017, Dssd: Deconvolutional single shot detector
[7]   An Anchor-Free Method Based on Feature Balancing and Refinement Network for Multiscale Ship Detection in SAR Images [J].
Fu, Jiamei ;
Sun, Xian ;
Wang, Zhirui ;
Fu, Kun .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2021, 59 (02) :1331-1344
[8]   Beyond Self-Attention: External Attention Using Two Linear Layers for Visual Tasks [J].
Guo, Meng-Hao ;
Liu, Zheng-Ning ;
Mu, Tai-Jiang ;
Hu, Shi-Min .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (05) :5436-5447
[9]  
Hong M., 2022, IEEE Geosci. Remote Sens. Lett., V19, P1
[10]  
Hu J, 2018, PROC CVPR IEEE, P7132, DOI [10.1109/TPAMI.2019.2913372, 10.1109/CVPR.2018.00745]