中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 先进控制与自动化团队
A Parallel Teacher for Synthetic-to-Real Domain Adaptation of Traffic Object Detection

文献类型：期刊论文

作者Wang, Jiangong5,6; Shen, Tianyu4; Tian, Yonglin6; Wang, Yutong6; Gou, Chao3; Wang, Xiao6; Yao, Fei2; Sun, Changyin1
刊名IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
出版日期2022-09-01
卷号7期号:3页码:441-455
关键词Object detection Feature extraction Data models Training Knowledge engineering Detectors Computational modeling Computer vision Unsupervised Domain Adaptation Teacher-student learning Traffic object detection
ISSN号2379-8858
DOI10.1109/TIV.2022.3197818
通讯作者Wang, Xiao(x.wang@ia.ac.cn)
英文摘要Large-scale synthetic traffic image datasets have been widely used to make compensate for the insufficient data in real world. However, the mismatch in domain distribution between synthetic datasets and real datasets hinders the application of the synthetic dataset in the actual vision system of intelligent vehicles. In this paper, we propose a novel synthetic-to-real domain adaptation method to settle the mismatch domain distribution from two aspects, i.e., data level and knowledge level. On the data level, a Style-Content Discriminated Data Recombination (SCD-DR) module is proposed, which decouples the style from content and recombines style and content from different domains to generate a hybrid domain as a transition between synthetic and real domains. On the knowledge level, a novel Iterative Cross-Domain Knowledge Transferring (ICD-KT) module including source knowledge learning, knowledge transferring and knowledge refining is designed, which achieves not only effective domain-invariant feature extraction, but also transfers the knowledge from labeled synthetic images to unlabeled actual images. Comprehensive experiments on public virtual and real dataset pairs demonstrate the effectiveness of our proposed synthetic-to-real domain adaptation approach in object detection of traffic scenes.
WOS关键词INTELLIGENT VEHICLES ; TRACKING ; VISION ; NETWORKS ; SYSTEMS ; IMAGES
资助项目Key-Area Research and Development Program of Guangdong Province[2020B090921003] ; National Natural Science Foundation of China[U1811463] ; Key Research and Development Program 2020 of Guangzhou[202007050002] ; Shenzhen Science and Technology Program[RCBS20200714114920272]
WOS研究方向Computer Science ; Engineering ; Transportation
语种英语
WOS记录号WOS:000873905600008
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
资助机构Key-Area Research and Development Program of Guangdong Province ; National Natural Science Foundation of China ; Key Research and Development Program 2020 of Guangzhou ; Shenzhen Science and Technology Program
源URL[http://ir.ia.ac.cn/handle/173211/50544]  
专题自动化研究所_复杂系统管理与控制国家重点实验室_先进控制与自动化团队
通讯作者Wang, Xiao
作者单位1.Anhui Univ, Sch Artificial Intelligence, Hefei 230039, Peoples R China
2.North Automat Controltechnol Inst, Taiyuan 030006, Peoples R China
3.Sun Yat Sen Univ, Sch Intelligent Syst Engn, Shenzhen Campus, Shenzhen 518107, Peoples R China
4.Beijing Normal Univ, Sch Artificial Intelligence, Beijing 100875, Peoples R China
5.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
6.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Wang, Jiangong,Shen, Tianyu,Tian, Yonglin,et al. A Parallel Teacher for Synthetic-to-Real Domain Adaptation of Traffic Object Detection[J]. IEEE TRANSACTIONS ON INTELLIGENT VEHICLES,2022,7(3):441-455.
APA Wang, Jiangong.,Shen, Tianyu.,Tian, Yonglin.,Wang, Yutong.,Gou, Chao.,...&Sun, Changyin.(2022).A Parallel Teacher for Synthetic-to-Real Domain Adaptation of Traffic Object Detection.IEEE TRANSACTIONS ON INTELLIGENT VEHICLES,7(3),441-455.
MLA Wang, Jiangong,et al."A Parallel Teacher for Synthetic-to-Real Domain Adaptation of Traffic Object Detection".IEEE TRANSACTIONS ON INTELLIGENT VEHICLES 7.3(2022):441-455.

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。