Correlation Filters Based on Multi-Expert and Game Theory for Visual Object Tracking

Ma, Songde; Zhao, Zhixian; Hou, Zhiqiang; Zhang, Lei; Yang, Xiaobao; Pu, Lei

doi:10.1109/tim.2022.3178482

Cited by 14 publications

(5 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Success Precision ULAST [41] 0.610 0.811 Spiking SiamFC++ [40] 0.644 0.854 DaHCF [42] 0.670 0.911 MEGTCF [44] 0.678 0.914 AFCSCF [43] 0.681 0.915 RPFormer [46] 0.682 0.891 DHPR [45] 0.691 0.916 ALT [47] 0.692 0.908 SiamTPN [37] 0…”

Section: Trackersmentioning

confidence: 99%

“…A  R  EAO  SiamSNN [25] 0.460 0.860 0.176 Siamese-RPN [26] 0.490 0.460 0.244 DHPR [45] 0.495 0.304 0.274 SiamFC [8] 0.503 0.585 0.188 MEGTCF [44] 0.505 0.314 0.278 UpdateNet [27] 0.518 0.454 0.244 DaSiamRPN+Att [28] 0.536 0.144 0.097 C-RPN [29] 0.550 0.320 0.273 BCS [31] 0.556 0.318 0.304 Spiking SiamFC++ [40] 0.556 0.445 0.255 Ours 0.559 0.323 0.304 RPFormer [46] 0.648 0.158 0.491 Trackers Success Precision ASTCA [38] 0.481 0.687 MRCF [20] 0.485 0.666 MEGTCF [44] 0.502 0.721 DHPR [45] 0.514 0.741 Spiking SiamFC++ [40] 0.578 0.744 SiamCAR [15] 0.614 0.760 SiamBAN [6] 0.631 0.833 SiamTPN [37] 0.636 0.823 UAST [50] 0.645 0.860 SiamGAT [39] 0.646 0.843 ALT [47] 0.652 0.871 Ours 0.656 0.873…”

Section: Trackersmentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Single Object Tracking With a Hybrid Approach: Temporal Convolutional Networks, Attention Mechanisms, and Spatial–Temporal Memory

Cheewaprakobkit,

Lin,

Shih

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Deep neural network-based tracking tasks have experienced significant advancements in recent years. However, these networks continue to face challenges in effectively adapting to appearance changes in both target and background, as well as linking objects after extended periods. The primary challenge in tracking lies in the frequent changes in a target's appearance throughout the tracking process, which can potentially reduce tracker robustness when faced with issues such as aspect ratio changes, occlusions, scale variations, and confusion from similar objects. To address this challenge, we propose a tracking architecture that combines a temporal convolutional network (TCN) and attention mechanism with spatial-temporal memory. The TCN component empowers the model to capture temporal dependencies, while the attention mechanism reduces computational complexity by focusing on crucial regions based on context. We leverage the target's historical information stored in the spatial-temporal memory network to guide the tracker in better adapting to target deformation. Our model attains a 63.0% average overlap (AO) on the GOT-10K dataset, a 71.9% success score (AUC) on OTB2015, a 65.6% success score (AUC) on UAV123, and achieves 56.9% accuracy on the VOT2018 dataset. These outcomes demonstrate the high effectiveness of our proposed tracker in tracking a single object.

show abstract

Section: Trackersmentioning

confidence: 99%

Section: Trackersmentioning

confidence: 99%

Enhancing Single Object Tracking With a Hybrid Approach: Temporal Convolutional Networks, Attention Mechanisms, and Spatial–Temporal Memory

Cheewaprakobkit,

Lin,

Shih

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…[44] dataset is one of the most authoritative benchmarks in the feld of object tracking, containing 100 completely labeled video sequences with 11 diferent challenge attributes, including illumination variation (IV), scale variation (SV), occlusion (OCC), deformation (DEF), motion blur (MB), fast motion (FM), inplane rotation (IPR), out-of-plane rotation (OPR), outof-view (OV), background clutters (BC), and low resolution (LR). We compared the proposed algorithms to 16 topperforming deep feature-based DCF trackers (DaHCF [16], AFCSCF [17], MEGTCF [52], DeepSTRCF [30], ECO [18], CFWCR [39], and C-COT [49]), handcrafted feature-based DCF trackers (SRECF [50], DRCF [33], AutoTrack [34], BACF [14], and SAMF [13]), and deep learning-based trackers (PrDiMP-18 [51], GradNet [53], DaSiamRPN [54], and SiamFC [38]). Figure 6 shows the proposed algorithm's tracking results compared to 16 other algorithms on the OTB2015 dataset.…”

Section: Qualitative Analysismentioning

confidence: 99%

“…Te UAV also employs the OPE method to evaluate the tracker, considering both success rate and precision. We compared our tracker with 10 leading-edge trackers, including DaHCF [16], MEGTCF [52], ASRCF [21], MCCT [22], ECO [18], MRCF [11], DRCF [33], STRCF [30], BACF [14], and SAMF [13], and the evaluation results are shown in Figure 12. Compared to the most recent state-of-the-art trackers (MRCF, MEGTCF, and DaHCF), our tracker surpasses the majority of its competitors, with the best DP score (75.3%) and the second highest AUC score (51.4%).…”

Section: Uav123 Datasetmentioning

confidence: 99%

Robust Visual Object Tracking Based on Feature Channel Weighting and Game Theory

Ma,

Zhao,

Hou

et al. 2023

International Journal of Intelligent Systems

Self Cite

View full text Add to dashboard Cite

Although the discriminative correlation filter- (DCF)-based tracker improves tracking performance, some object representation issues can still be further optimized. On the one hand, the DCF tracker’s deep convolutional features contain many noisy channels, and assigning the same weights to multiple channels cannot distinguish the importance of different channels. On the other hand, a simple weighted fusion approach cannot fully utilize the benefits of different feature types. We propose a visual object tracking algorithm based on adaptive channel weighting and feature game fusion to solve these problems. In this study, an adaptive channel weighting strategy is designed to assign suitable weights to each channel based on the average energy ratio of the target and background regions in the feature channels and prune the channels with low weights to improve feature robustness and reduce computational complexity. Simultaneously, the game theory concept is introduced in the multifeature fusion. The handcrafted features are combined with shallow and deep convolutional features according to feature complementarity. Then, the two combined features are seen as two sides of the game, continuously gamed during the tracking process to generate a feature model with a higher representation capacity. Extensive experiments are conducted on four mainstream visual tracking benchmark datasets, including OTB2015, VOT2018, LaSOT, and UAV123. The experimental results show that the proposed algorithm performs outstandingly compared to the state-of-the-art trackers.

show abstract

“…Although effective, its performance is contingent on the accuracy of the system model, which is not always guaranteed. Adaptive filtering methods like the least mean squares and Recursive least squares algorithms also offer promise but require careful selection of suitable parameters for optimal performance [12].…”

Section: Introductionmentioning

confidence: 99%

Laser self-mixing interference displacement signal filtering method based on empirical mode decomposition and wavelet threshold

Guo,

Wang

2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Objective:In laser self-mixing interferometry displacement measurement, noise interference has a significant impact on the measurement results. To improve measurement accuracy, this paper proposes a filtering method that combines empirical mode decomposition (EMD) with wavelet thresholding.Method:First, the signal is decomposed into several intrinsic mode functions (IMFs) using EMD. Then, wavelet thresholding is applied to each IMF. Subsequently, the processed IMFs are reconstructed to achieve signal filtering. Finally, by integrating the principles of interpolation and fringe counting, the reconstructed displacement signal is recovered, realizing accurate displacement measurement.Result:This paper presents comprehensive simulation analyses and experimental validations for the proposed method. The accuracy of the displacement recovery is quantitatively evaluated using the absolute error and standard error, comparing the recovered displacement signal with the actual displacement.Conclusion:The experimental results demonstrate that the laser self-mixing interferometry displacement signal filtering method based on EMD and wavelet thresholding has high accuracy.

show abstract

Correlation Filters Based on Multi-Expert and Game Theory for Visual Object Tracking

Cited by 14 publications

References 50 publications

Enhancing Single Object Tracking With a Hybrid Approach: Temporal Convolutional Networks, Attention Mechanisms, and Spatial–Temporal Memory

Enhancing Single Object Tracking With a Hybrid Approach: Temporal Convolutional Networks, Attention Mechanisms, and Spatial–Temporal Memory

Robust Visual Object Tracking Based on Feature Channel Weighting and Game Theory

Laser self-mixing interference displacement signal filtering method based on empirical mode decomposition and wavelet threshold

Contact Info

Product

Resources

About