期刊
NEUROCOMPUTING
卷 513, 期 -, 页码 204-214出版社
ELSEVIER
DOI: 10.1016/j.neucom.2022.09.028
关键词
Visual tracking; Siamese network; Transformer; Similarity measuring
资金
- National Natural Science Foundation of China (NSFC)
- Defense Industrial Technology Devel- opment Program
- [81671787]
- [JCKY2016208B001]
- [LabSOMP-2018-03]
The paper proposes a similarity measurement method called Dynamic Cross-Attention (DCA), which utilizes Transformer encoders to explore feature interdependency and redesigns each part of the Siamese network, resulting in improved tracking performance.
Most of the existing Siamese tracking methods follow the overall framework of SiamRPN, adopting its general network architecture and the local and linear cross-correlation operation to integrate search and template features, which restricts the introduction of more sophisticated structures for expressive appearance representation as well as the further improvements on tracking performance. Motivated by the recent progresses in vision Transformer and MLP, we first explore to accomplish a global, nonlinear and scale-invariant similarity measuring manner called Dynamic Cross-Attention (DCA). Specifically, template features are first decomposed along the spatial and channel dimension and then the Transformer Encoders are applied to adaptively excavate the long-range feature interdependency, pro-ducing reinforced kernels. As the kernels are successively multiplied to the search feature map, similarity scores between all the pixels on feature maps are estimated at once while the spatial scale of search fea-tures remains constant. Furthermore, we redesign each part of our Siamese network to further remedy the framework limitation with the assistant of DCA. Comprehensive experimental results on large-scale benchmarks indicate that our Siamese method realizes the efficient feature extraction, aggregation, refinement and interaction, outperforming state-of-the-art trackers. (c) 2022 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据