Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Object Tracking

Xu, Tianyang; Feng, Zhen-Hua; Wu, Xiao-Jun; Kittler, Josef

doi:10.1109/tip.2019.2919201

Cited by 360 publications

(211 citation statements)

References 79 publications

Supporting

Mentioning

211

Contrasting

Order By: Relevance

“…A common characterisation of the above approaches is that they are all based on a fixed spatial regularisation pattern, for example, a predefined mask or weighting function. To achieve adaptive spatial regularisation, LADCF [83] embeds dynamic spatial feature selection in the filter learning stage. Thanks to this innovation it has achieved the best results in the public VOT2018 dataset [34].…”

Section: Related Workmentioning

confidence: 99%

“…We evaluated the proposed method on several wellknown benchmarks, including OTB2013/OTB2015 [81,82], VOT2017/VOT2018 [33,34] and TrackingNet Test dataset [55], and compared it with a number of state-of-theart trackers, such as VITAL [68], MetaT [58], ECO [13], MCPF [89], CREST [67], BACF [31], CFNet [73], CACF [54], ACFN [11], CSRDCF [49], C-COT [51], Staple [4], SiamFC [5], SRDCF [15], KCF [27], SAMF [41], DSST [16] and other advanced trackers in VOT challenges, i.e., CFCF [23], CFWCR [25], LSART [69], UPDT [6], SiamRPN [91], MFT [34] and LADCF [83].…”

Section: Implementation and Evaluation Settingsmentioning

confidence: 99%

See 1 more Smart Citation

Joint Group Feature Selection and Discriminative Filter Learning for Robust Visual Object Tracking

Feng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

Self Cite

197

118

View full text Add to dashboard Cite

We propose a new Group Feature Selection method for Discriminative Correlation Filters (GFS-DCF) based visual object tracking. The key innovation of the proposed method is to perform group feature selection across both channel and spatial dimensions, thus to pinpoint the structural relevance of multi-channel features to the filtering system. In contrast to the widely used spatial regularisation or feature selection methods, to the best of our knowledge, this is the first time that channel selection has been advocated for DCF-based tracking. We demonstrate that our GFS-DCF method is able to significantly improve the performance of a DCF tracker equipped with deep neural network features. In addition, our GFS-DCF enables joint feature selection and filter learning, achieving enhanced discrimination and interpretability of the learned filters.To further improve the performance, we adaptively integrate historical information by constraining filters to be smooth across temporal frames, using an efficient lowrank approximation. By design, specific temporal-spatialchannel configurations are dynamically learned in the tracking process, highlighting the relevant features, and alleviating the performance degrading impact of less discriminative representations and reducing information redundancy. The experimental results obtained on OTB2013, OTB2015, VOT2017, VOT2018 and TrackingNet demonstrate the merits of our GFS-DCF and its superiority over the state-of-the-art trackers. The code is publicly available at https://github.com/XU-TIANYANG/ GFS-DCF.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Implementation and Evaluation Settingsmentioning

confidence: 99%

Joint Group Feature Selection and Discriminative Filter Learning for Robust Visual Object Tracking

Feng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

Self Cite

197

118

View full text Add to dashboard Cite

show abstract

“…Although SATIN ranks the 15th with an EAO score of 0.294 in the baseline experiment, it achieves a third-best performance with the EAO score of 0.289 in the real-time experiment. Compared with the LADCF tracker [71], which delivers the best EAO score of 0.389 in the baseline experiment, SATIN outperforms it by 22.3% in the real-time experiment. SiamRPN [25] obtains the best EAO score of 0.383 in the real-time experiment 2 .…”

Section: Experiments On Votmentioning

confidence: 96%

Siamese attentional keypoint network for high performance visual tracking

Gao

Yuan

Wang

et al. 2020

Knowledge-Based Systems

View full text Add to dashboard Cite

Visual tracking is one of the most fundamental topics in computer vision. Numerous tracking approaches based on discriminative correlation filters or Siamese convolutional networks have attained remarkable performance over the past decade. However, it is still commonly recognized as an open research problem to develop robust and effective trackers which can achieve satisfying performance with high computational and memory storage efficiency in real-world scenarios. In this paper, we investigate the impacts of three main aspects of visual tracking, i.e., the backbone network, the attentional mechanism, and the detection component, and propose a Siamese Attentional Keypoint Network, dubbed SATIN, for efficient tracking and accurate localization. Firstly, a new Siamese lightweight hourglass network is specially designed for visual tracking. It takes advantage of the benefits of the repeated bottom-up and top-down inference to capture more global and local contextual information at multiple scales. Secondly, a novel cross-attentional module is utilized to leverage both channel-wise and spatial intermediate attentional information, which can enhance both discriminative and localization capabilities of feature maps. Thirdly, a keypoints detection approach is invented to trace any target object by detecting the top-left corner point, the centroid point, and the bottom-right corner point of its bounding box. Therefore, our SATIN tracker not only has a strong capability to learn more effective object representations, but also is computational and memory storage efficiency, either during the training or testing stages. To the best of our knowledge, we are the first to propose this approach. Without bells and whistles, experimental results demonstrate that our approach achieves state-of-the-art performance on several recent benchmark datasets, at a speed far exceeding 27 frames per second. with real-time speed are widely pursued in numerous applications, including video surveillance, autonomous driving, pose estimation, and human-computer interfaces. Although considerable progress has been made by many tracking approaches in the last decade, visual tracking is still a challenging problem due to multiple negative scenarios such as occlusions, fast motions, scale variations, and background clutters [1,2,3,4].Most recent tracking approaches are developed based on two successful frameworks. The first one is the discriminative correlation filter (DCF). Thanks to the circular shifting of training samples and fast learning correlation filters in the Fourier frequency domain, DCF demonstrates superior computational efficiency and fairly good tracking accuracy and has attracted increasing attention since MOSSE [5]first exploited it. Later, in order to overcome different kinds of challenges and achieve competitive tracking performance, the advancement in DCF-based tracking is focused on the use of kernel functions [6], motion information [7,8], multi-dimensional features [9,10], multi-scale estimation [11,12], boundary effects allevi...

show abstract

“…The entire set of circulant shifts satisfies the convolution theorem, simplifying the computation via element-wise multiplication and division in the frequency space, rather than calculating the product or inverse of training matrices. These advantages are further strengthened by incorporating additional regularisations and constraints [11,12]. Apart from the improvements achieved by sophisticated mathematical formulations, advanced DCF trackers tend to employ powerful deep CNN features for boosting the performance [13,14,15].…”

Section: Introductionmentioning

confidence: 99%

An accelerated correlation filter tracker

Xu¹,

Feng²,

Wu³

et al. 2020

Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Recent visual object tracking methods have witnessed a continuous improvement in the state-of-the-art with the development of efficient discriminative correlation filters (DCF) and robust deep neural network features. Despite the outstanding performance achieved by the above combination, existing advanced trackers suffer from the burden of high computational complexity of the deep feature extraction and online model learning. We propose an accelerated ADMM optimisation method obtained by adding a momentum to the optimisation sequence iterates, and by relaxing the impact of the error between DCF parameters and their norm. The proposed optimisation method is applied to an innovative formulation of the DCF design, which seeks the most discriminative spatially regularised feature channels. A further speed up is achieved by an adaptive initialisation of the filter optimisation process. The significantly increased convergence of the DCF filter is demonstrated by establishing the optimisation process equivalence with a continuous dynamical system for which the convergence properties can readily be derived. The experimental results obtained on several well-known benchmarking datasets demonstrate the efficiency and robustness of the proposed ACFT method, with a tracking accuracy comparable to the start-of-the-art trackers.

show abstract

Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Object Tracking

Cited by 360 publications

References 79 publications

Joint Group Feature Selection and Discriminative Filter Learning for Robust Visual Object Tracking

Joint Group Feature Selection and Discriminative Filter Learning for Robust Visual Object Tracking

Siamese attentional keypoint network for high performance visual tracking

An accelerated correlation filter tracker

Contact Info

Product

Resources

About