Adaptive multiscale feature for object detection | |
Yu, Xiaoyong2,3; Wu, Siyuan3; Lu, Xiaoqiang3; Gao, Guilong1 | |
刊名 | Neurocomputing |
2021-08-18 | |
卷号 | 449页码:146-158 |
关键词 | Object detection Classification network Backbone network Multiscale feature Feature fusion Adaptation Anchor Anchor-free |
ISSN号 | 09252312;18728286 |
DOI | 10.1016/j.neucom.2021.04.002 |
产权排序 | 2 |
英文摘要 | In object detection, multiscale features are necessary to deal with objects with different size. Using Feature Pyramid Network (FPN) as the backbone network is a very popular paradigm in existing object detectors, we call this paradigm FPN+. However, feature fusion of FPN is insufficient to express object of similar size but different appearance due to the unidirectional feature fusion. We motivate and present Adaptive Multiscale Feature (AMF), a new multiscale feature fusion method with bidirectional feature fusion, using to solve the one-direction fusion of FPN. AMF module fuses multiscale features from two aspects: (1) shattering features by the way of CLSM; (2) fusing features by the way of channel-wise attention. The proposed AMF improves the expression ability of multiscale features of the backbone network, and effectively improves the performance of the object detector. The proposed feature fusion method can be added to all object detector, such as the one-stage detector, the two-stage detector, anchor-based detector and anchor-free based detector. Experimental results on the COCO 2014 dataset show that the proposed AMF module performs the popular FPN based detector. Whether anchored-free based detectors or anchored based detectors, performance of detector can be improved through AMF. And the resulting best model can achieve a very competitive mAP on COCO 2014 dataset. © 2021 Elsevier B.V. |
语种 | 英语 |
出版者 | Elsevier B.V. |
内容类型 | 期刊论文 |
源URL | [http://ir.opt.ac.cn/handle/181661/94711] |
专题 | 西安光学精密机械研究所_光学影像学习与分析中心 |
通讯作者 | Wu, Siyuan |
作者单位 | 1.Key Laboratory of Ultra-fast Photoelectric Diagnostics Technology, Xi'an Institute of Optics and Precision Mechanics (XIOPM), Chinese Academy of Sciences (CAS), Xi'an; Shaanxi; 710119, China 2.University of Chinese Academy of Sciences, Beijing; 100049, China; 3.Key Laboratory of Spectral Imaging Technology CAS, Xi'an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi'an; Shaanxi; 710119, China; |
推荐引用方式 GB/T 7714 | Yu, Xiaoyong,Wu, Siyuan,Lu, Xiaoqiang,et al. Adaptive multiscale feature for object detection[J]. Neurocomputing,2021,449:146-158. |
APA | Yu, Xiaoyong,Wu, Siyuan,Lu, Xiaoqiang,&Gao, Guilong.(2021).Adaptive multiscale feature for object detection.Neurocomputing,449,146-158. |
MLA | Yu, Xiaoyong,et al."Adaptive multiscale feature for object detection".Neurocomputing 449(2021):146-158. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论