Siamese Visual Tracking With Residual Fusion Learning

doi:10.1109/access.2021.3134066

CORC > 长春光学精密机械与物理研究所 > 中国科学院长春光学精密机械与物理研究所

	Siamese Visual Tracking With Residual Fusion Learning
	X. L. Sun; G. L. Han and L. H. Guo
刊名	Ieee Access
	2022
卷号	10 页码:88421-88433
ISSN号	2169-3536
DOI	10.1109/access.2021.3134066
英文摘要	Multi-stage feature fusion is pretty effective for deep Siamese trackers to promote tracking performance. Unfortunately, conventional fusion approaches, such as weighted average, are so simple that they are inappropriate to combine the features with diverse characteristics. In addition, the fusion module is generally optimized along with Siamese network module, which may result in the performance degradation of the whole tracker. In this paper, we propose a novel feature fusion network for Siamese tracker by exploiting the expression capacity of residual fusion learning (SiamRFL). Specifically, the network employs the deep-layer features as direct input to semantically recognize the object from background, and refines the object state with local detail patterns by exploring the shallow-layer features through residual channel. The classification and the regression features can be fused respectively by deploying multiple fusion units. To avoid the degradation problem, we also present an ensemble training framework for our tracker, in which different loss functions are introduced to individually optimize the Siamese and the fusion modules. Compared to the baseline SiamRPN++ tracker, the proposed tracker achieves favorable gains by 0.696 -> 0.709, 0.285 -> 0.308, 0.603 -> 0.624, 0.496 -> 0.520 and 0.517 -> 0.559 on OTB100, VOT2019, UAV123, LaSOT and GOT1Ok datasets, outperforming other approaches by an obvious margin.
URL标识	查看原文
语种	英语
内容类型	期刊论文
源URL	[http://ir.ciomp.ac.cn/handle/181722/67075]
专题	中国科学院长春光学精密机械与物理研究所
推荐引用方式 GB/T 7714	X. L. Sun,G. L. Han and L. H. Guo. Siamese Visual Tracking With Residual Fusion Learning[J]. Ieee Access,2022,10:88421-88433.
APA	X. L. Sun,&G. L. Han and L. H. Guo.(2022).Siamese Visual Tracking With Residual Fusion Learning.Ieee Access,10,88421-88433.
MLA	X. L. Sun,et al."Siamese Visual Tracking With Residual Fusion Learning".Ieee Access 10(2022):88421-88433.