Regularizing deep networks with label geometry for accurate object localization on small training datasets
文献类型：期刊论文
| 作者 | Wang, Xiaolian2,3 ; Hu, Xiyuan4; Chen, Chen2,3; Peng, Silong1,2,3
|
| 刊名 | PATTERN RECOGNITION LETTERS
 |
| 出版日期 | 2022-02-01 |
| 卷号 | 154页码:53-59 |
| 关键词 | Object detection Object localization Label geometry Box evolution Small dataset Human-machine interaction |
| ISSN号 | 0167-8655 |
| DOI | 10.1016/j.patrec.2022.01.004 |
| 通讯作者 | Hu, Xiyuan(huxy@njust.edu.cn) |
| 英文摘要 | Localization is a critical subtask in object detection, which is closely related to spatial information of objects. Most current detectors simply rely on the fitting ability of deep neural networks to regress towards numerical targets such as coordinates of object boxes. Training deep networks for sufficient fitting ability requires a large number of annotations that are expensive to obtain. In this work, we fully exploit limited annotations by extracting label geometry to improve localization performance on small datasets. We generate distance transform of bounding box edges according to localization labels, with which we supervise intermediate outputs of networks pixel by pixel to reconstruct object geometry for localization. Distance transform is sensitive to box edges and provides geometric gradients flowing into boundaries. We learn such gradients to enhance geometric-aware features through a coupled training with regression, and use it to refine regressed boxes in an evolutionary manner in inference. Extensive experiments are implemented to demonstrate the effectiveness of our method. Our method can be applied in applications that required human-machine interaction, such as the driver-assistance system in autonomous driving, by providing accurate detections to assist humans in making better decisions.(c) 2022 Elsevier B.V. All rights reserved. |
| 资助项目 | National Key R&D Program of China[2021YFF0602101] ; National Science Founda-tion of China[NSFC 61906194] |
| WOS研究方向 | Computer Science |
| 语种 | 英语 |
| WOS记录号 | WOS:000783134200006 |
| 出版者 | ELSEVIER |
| 资助机构 | National Key R&D Program of China ; National Science Founda-tion of China |
| 源URL | [http://ir.ia.ac.cn/handle/173211/48347]  |
| 专题 | 自动化研究所_智能制造技术与系统研究中心_多维数据分析团队 |
| 通讯作者 | Hu, Xiyuan |
| 作者单位 | 1.Beijing ViSyst Corp Ltd, Beijing, Peoples R China 2.Chinese Acad Sci, Inst Automat, Beijing, Peoples R China 3.Univ Chinese Acad Sci, Beijing, Peoples R China 4.Nanjing Univ Sci & Technol, Nanjing, Jiangsu, Peoples R China |
| 推荐引用方式 GB/T 7714 | Wang, Xiaolian,Hu, Xiyuan,Chen, Chen,et al. Regularizing deep networks with label geometry for accurate object localization on small training datasets[J]. PATTERN RECOGNITION LETTERS,2022,154:53-59. |
| APA | Wang, Xiaolian,Hu, Xiyuan,Chen, Chen,&Peng, Silong.(2022).Regularizing deep networks with label geometry for accurate object localization on small training datasets.PATTERN RECOGNITION LETTERS,154,53-59. |
| MLA | Wang, Xiaolian,et al."Regularizing deep networks with label geometry for accurate object localization on small training datasets".PATTERN RECOGNITION LETTERS 154(2022):53-59. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
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。