Antidecay LSTM for Siamese Tracking With Adversarial Learning

doi:10.1109/TNNLS.2020.3018025

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 图像与视频分析团队

	Antidecay LSTM for Siamese Tracking With Adversarial Learning
	Zhao, Fei 4,5; Zhang, Ting 6; Wu, Yi 1; Tang, Ming 5,7; Wang, Jinqiao 2,3,5
刊名	IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
	2021-10-01
卷号	32 期号:10 页码:4475-4489
关键词	Target tracking Feature extraction Computer architecture Visualization Training Task analysis Adversarial learning deep learning long short-term memory (LSTM) visual tracking
ISSN号	2162-237X
DOI	10.1109/TNNLS.2020.3018025
通讯作者	Wang, Jinqiao(jqwang@nlpr.ia.ac.cn)
英文摘要	Visual tracking is one of the fundamental tasks in computer vision with many challenges, and it is mainly due to the changes in the target's appearance in temporal and spatial domains. Recently, numerous trackers model the appearance of the targets in the spatial domain well by utilizing deep convolutional features. However, most of these CNN-based trackers only take the appearance variations between two consecutive frames in a video sequence into consideration. Besides, some trackers model the appearance of the targets in the long term by applying RNN, but the decay of the target's features degrades the tracking performance. In this article, we propose the antidecay long short-term memory (AD-LSTM) for the Siamese tracking. Especially, we extend the architecture of the standard LSTM in two aspects for the visual tracking task. First, we replace all of the fully connected layers with convolutional layers to extract the features with spatial structure. Second, we improve the architecture of the cell unit. In this way, the information of the target appearance can flow through the AD-LSTM without decay as long as possible in the temporal domain. Meanwhile, since there is no ground truth for the feature maps generated by the AD-LSTM, we propose an adversarial learning algorithm to optimize the AD-LSTM. With the help of adversarial learning, the Siamese network can generate the response maps more accurately, and the AD-LSTM can generate the feature maps of the target more robustly. The experimental results show that our tracker performs favorably against the state-of-the-art trackers on six challenging benchmarks: OTB-100, TC-128, VOT2016, VOT2017, GOT-10k, and TrackingNet.
资助项目	Research and Development Projects in the Key Areas of Guangdong Province[2020B010165001] ; National Natural Science Foundation of China[61772527] ; National Natural Science Foundation of China[61976210] ; National Natural Science Foundation of China[61806200] ; National Natural Science Foundation of China[61702510] ; National Natural Science Foundation of China[61876086]
WOS关键词	OBJECT TRACKING ; NETWORKS
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000704111000019
资助机构	Research and Development Projects in the Key Areas of Guangdong Province ; National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/46200]
专题	自动化研究所_模式识别国家重点实验室_图像与视频分析团队
通讯作者	Wang, Jinqiao
作者单位	1.Wormpex AI Res, Bellevue, WA 98004 USA 2.ObjectEye Inc, Beijing 100080, Peoples R China 3.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 4.Alibaba Grp, Hangzhou 311100, Peoples R China 5.Chinese Acad Sci, Natl Lab Pattern Recognit, Inst Automat, Beijing 100190, Peoples R China 6.Ctr China Natl Elect Import & Export Corp CEIEC, Ctr Res & Dev, Beijing 100036, Peoples R China 7.Shenzhen Infinova Ltd, Shenzhen 518110, Peoples R China
推荐引用方式 GB/T 7714	Zhao, Fei,Zhang, Ting,Wu, Yi,et al. Antidecay LSTM for Siamese Tracking With Adversarial Learning[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2021,32(10):4475-4489.
APA	Zhao, Fei,Zhang, Ting,Wu, Yi,Tang, Ming,&Wang, Jinqiao.(2021).Antidecay LSTM for Siamese Tracking With Adversarial Learning.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,32(10),4475-4489.
MLA	Zhao, Fei,et al."Antidecay LSTM for Siamese Tracking With Adversarial Learning".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 32.10(2021):4475-4489.