Knowledge Commons of Institute of Automation,CAS
SOTVerse: A User-Defined Task Space of Single Object Tracking | |
Shiyu, Hu1,2![]() ![]() | |
发表期刊 | International Journal of Computer Vision
![]() |
ISSN | 0920-5691 |
2023-10 | |
卷号 | 132期号:3页码:1-59 |
通讯作者 | Zhao, Xin(xzhao@nlpr.ia.ac.cn) ; Huang, Kaiqi(kqhuang@nlpr.ia.ac.cn) |
摘要 | Single object tracking (SOT) research falls into a cycle—trackers perform well on most benchmarks but quickly fail in challenging scenarios, causing researchers to doubt the insufficient data content and take more effort to construct larger datasets with more challenging situations. However, inefficient data utilization and limited evaluation methods more seriously hinder SOT research. The former causes existing datasets can not be exploited comprehensively, while the latter neglects challenging factors in the evaluation process. In this article, we systematize the representative benchmarks and form a single object tracking metaverse (SOTVerse)—a user-defined SOT task space to break through the bottleneck. We first propose a 3E Paradigm to describe tasks by three components (i.e., environment, evaluation, and executor). Then, we summarize task characteristics, clarify the organization standards, and construct SOTVerse with 12.56 million frames. Specifically, SOTVerse automatically labels challenging factors per frame, allowing users to generate user-defined spaces efficiently via construction rules. Besides, SOTVerse provides two mechanisms with new indicators and successfully evaluates trackers under various subtasks. Consequently, SOTVerse first provides a strategy to improve resource utilization in the computer vision area, making research more standardized. The SOTVerse, toolkit, evaluation server, and results are available at http://metaverse.aitestunion.com. |
关键词 | Single object tracking Experimental environment Evaluation system Performance analysis |
DOI | 10.1007/s11263-023-01908-5 |
关键词[WOS] | PERFORMANCE ; TIME |
收录类别 | SCI |
语种 | 英语 |
资助项目 | Youth Innovation Promotion Association of the Chinese Academy of Sciences |
项目资助者 | Youth Innovation Promotion Association of the Chinese Academy of Sciences |
WOS研究方向 | Computer Science |
WOS类目 | Computer Science, Artificial Intelligence |
WOS记录号 | WOS:001176990700015 |
出版者 | SPRINGER |
七大方向——子方向分类 | 目标检测、跟踪与识别 |
国重实验室规划方向分类 | 智能能力评估 |
是否有论文关联数据集需要存交 | 否 |
引用统计 | |
文献类型 | 期刊论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/54542 |
专题 | 复杂系统认知与决策实验室_智能系统与工程 |
通讯作者 | Xin, Zhao; Kaiqi Huang |
作者单位 | 1.School of Artificial Intelligence, University of Chinese Academy of Sciences, 2.Institute of Automation, Chinese Academy of Sciences 3.Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences |
第一作者单位 | 中国科学院自动化研究所 |
通讯作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Shiyu, Hu,Xin, Zhao,Kaiqi Huang. SOTVerse: A User-Defined Task Space of Single Object Tracking[J]. International Journal of Computer Vision,2023,132(3):1-59. |
APA | Shiyu, Hu,Xin, Zhao,&Kaiqi Huang.(2023).SOTVerse: A User-Defined Task Space of Single Object Tracking.International Journal of Computer Vision,132(3),1-59. |
MLA | Shiyu, Hu,et al."SOTVerse: A User-Defined Task Space of Single Object Tracking".International Journal of Computer Vision 132.3(2023):1-59. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
SOTVerse.pdf(53048KB) | 期刊论文 | 作者接受稿 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
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