Filters








6,492 Hits in 3.6 sec

Adaptive Exploitation of Pre-trained Deep Convolutional Neural Networks for Robust Visual Tracking [article]

Seyed Mojtaba Marvasti-Zadeh, Hossein Ghanei-Yakhdan, Shohreh Kasaei
2020 arXiv   pre-print
Although many of those trackers utilize the feature maps from pre-trained convolutional neural networks (CNNs), the effects of selecting different models and exploiting various combinations of their feature  ...  Second, with the aid of analysis results as attribute dictionaries, adaptive exploitation of deep features is proposed to improve the accuracy and robustness of visual trackers regarding video characteristics  ...  Conflict of interest All authors declare that they have no conflict of interest.  ... 
arXiv:2008.13015v2 fatcat:r5b6kd3gfvcadf7xbviaxpd4zi

An Analytical Review on Some Recent Advances in Deep Learning Object Tracking Approaches

K. Nani Kumar, M. James Stephen, P. V. G. D. Prasad Reddy, Andhra University
2020 International Journal of Engineering Research and  
This paper presents a detailed review on some of the recent advances in Deep Learning Based Visual Object Tracking Approaches from a wide variety of algorithms often cited in research literature.  ...  Visual Object tracking in real world, real time application scenarios is a complex problem, therefore, it remains a most active area of research in computer vision.  ...  In this method the deep network to control actions is pre-trained using various training sequences and fine-tuned during tracking for online adaptation to target and background changes.  ... 
doi:10.17577/ijertv9is010309 fatcat:e7wny2gl35cuvfxrcfec3zxn7y

Deep Learning for Visual Tracking: A Comprehensive Survey [article]

Seyed Mojtaba Marvasti-Zadeh, Li Cheng, Hossein Ghanei-Yakhdan, and Shohreh Kasaei
2019 arXiv   pre-print
visual tracking, network objective, network output, and the exploitation of correlation filter advantages.  ...  First, the fundamental characteristics, primary motivations, and contributions of DL-based methods are summarized from six key aspects of: network architecture, network exploitation, network training for  ...  Kamal Nasrollahi (Visual Analysis of People Lab (VAP), Aalborg University) for his beneficial comments.  ... 
arXiv:1912.00535v1 fatcat:v5ikqi2cpbblhgtkiu6z6l5anq

Efficient Scale Estimation Methods using Lightweight Deep Convolutional Neural Networks for Visual Tracking [article]

Seyed Mojtaba Marvasti-Zadeh, Hossein Ghanei-Yakhdan, Shohreh Kasaei
2020 arXiv   pre-print
Although a wide range of recent DCF-based methods exploit the features that are extracted from deep convolutional neural networks (CNNs) in their translation model, the scale of the visual target is still  ...  Whereas the exploitation of CNNs imposes a high computational burden, this paper exploits pre-trained lightweight CNNs models to propose two efficient scale estimation methods, which not only improve the  ...  Although deep neural networks have led to many significant advances in visual tracking methods (e.g., robust target representation), the state-of-theart DCF-based visual trackers still exploit hand-crafted  ... 
arXiv:2004.02933v2 fatcat:p3sqisjjpvhbrcyzceknelefnm

Deep Learning Trackers Review and Challenge

Yongxiang Gu, Beijing Chen, Xu Cheng, Yifeng Zhang, Jingang Shi
2019 Journal of Information Hiding and Privacy Protection  
. (4) The deep visual trackers using end-to-end networks usually perform better than the trackers merely using feature extraction networks. (5) For visual tracking, the most suitable network training method  ...  We conclude that: (1) The usage of the convolutional neural network (CNN) model could significantly improve the tracking performance. (2) The trackers with deep features perform much better than those  ...  Network training The network training is also a critical issue for developing a robust deep learning based tracker, which may be used to transfer visual prior, online learning, or both.  ... 
doi:10.32604/jihpp.2019.05938 fatcat:z2kq47sl25fz7fykzdhfozysge

Deep Siamese Networks toward Robust Visual Tracking [chapter]

Mustansar Fiaz, Arif Mahmood, Soon Ki Jung
2019 Visual Object Tracking in the Deep Neural Networks Era [Working Title]  
Recently, Siamese neural networks have been widely used in visual object tracking to leverage the template matching mechanism.  ...  Whereas in the early merge architecture, inputs are combined at the start of the network and a unified data stream is processed by a single convolutional neural network.  ...  Deep Siamese Networks toward Robust Visual Tracking DOI: http://dx.doi.org/10.5772/intechopen.86235 Deep Siamese Networks toward Robust Visual Tracking DOI: http://dx.doi.org/10.5772/intechopen.86235  ... 
doi:10.5772/intechopen.86235 fatcat:rgz7z6ldujczbefuznzhfetgsm

Effective Fusion of Deep Multitasking Representations for Robust Visual Tracking [article]

Seyed Mojtaba Marvasti-Zadeh, Hossein Ghanei-Yakhdan, Shohreh Kasaei, Kamal Nasrollahi, Thomas B. Moeslund
2021 arXiv   pre-print
terms of precision and robustness of visual tracking.  ...  This paper aims to evaluate the performance of twelve state-of-the-art ResNet-based FENs in a DCF-based framework to determine the best for visual tracking purposes.  ...  Compliance with Ethical Standards (Conflict of Interest): All authors declare that they have no conflict of interest.  ... 
arXiv:2004.01382v2 fatcat:hsns3y46g5c7vdghlmgbzlniji

Target-Aware Deep Tracking

Xin Li, Chao Ma, Baoyuan Wu, Zhenyu He, Ming-Hsuan Yang
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Existing deep trackers mainly use convolutional neural networks pre-trained for the generic object recognition task for representations.  ...  Despite demonstrated successes for numerous vision tasks, the contributions of using pre-trained deep features for visual tracking are not as significant as that for object recognition.  ...  We narrow the gap between the pre-trained deep models and target objects of arbitrary forms for visual tracking. • We integrate the target-aware features with a Siamese matching network for visual tracking  ... 
doi:10.1109/cvpr.2019.00146 dblp:conf/cvpr/Li0WH019 fatcat:hjod2ekhzbfjdonfhfxi6iekha

Beyond Background-Aware Correlation Filters: Adaptive Context Modeling by Hand-Crafted and Deep RGB Features for Visual Tracking [article]

Seyed Mojtaba Marvasti-Zadeh, Hossein Ghanei-Yakhdan, Shohreh Kasaei
2021 arXiv   pre-print
Comprehensive experimental results demonstrate that the proposed adaptive method clearly outperforms the accuracy and robustness of visual target tracking compared to the state-of-the-art methods on the  ...  neural network (CNN) feature maps.  ...  Compliance with Ethical Standards: All authors declare that they have no conflict of interest.  ... 
arXiv:2004.02932v2 fatcat:dbjgzsequvcpxgq2qlilv6rmbi

Online Visual Tracking with One-Shot Context-Aware Domain Adaptation [article]

Hossein Kashiani, Amir Abbas Hamidi Imani, Shahriar Baradaran Shokouhi, Ahmad Ayatollahi
2021 arXiv   pre-print
The domain adaptation approach is backboned with only an off-the-shelf deep model.  ...  Online learning policy makes visual trackers more robust against different distortions through learning domain-specific cues.  ...  Similarly, [21] and [20] pre-train Siamese networks with different structures for online tracking.  ... 
arXiv:2008.09891v2 fatcat:2o57dm6rozggrin6pun6fwwtpq

STCT: Sequentially Training Convolutional Networks for Visual Tracking

Lijun Wang, Wanli Ouyang, Xiaogang Wang, Huchuan Lu
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we propose a sequential training method for convolutional neural networks (CNNs) to effectively transfer pre-trained deep features for online applications.  ...  Due to the limited amount of training samples, finetuning pre-trained deep models online is prone to overfitting.  ...  In order to address the above issue, we propose a sequential training method for CNNs to effectively transfer pre-trained deep features for online visual tracking.  ... 
doi:10.1109/cvpr.2016.153 dblp:conf/cvpr/WangOWL16 fatcat:c7rbqdxnvfdhdc7g4zxx2zviou

Learning Spatial-Aware Regressions for Visual Tracking

Chong Sun, Dong Wang, Huchuan Lu, Ming-Hsuan Yang
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
In this paper, we analyze the spatial information of deep features, and propose two complementary regressions for robust visual tracking.  ...  Distance transform pooling is further exploited to determine the effectiveness of each output channel of the convolution layer.  ...  This paper is partially supported by the Natural Science Foundation of China #61725202, #61502070, #61472060, NSF CAREER (No. 1149783), gifts from Adobe, Toyota, Panasonic, Samsung, NEC, Verisk and Nvidia  ... 
doi:10.1109/cvpr.2018.00934 dblp:conf/cvpr/Sun0L018a fatcat:2maquvcqg5ahjm6nmb4tq2wm3e

Siamese Visual Object Tracking: A Survey

Milan Ondrasovic, Peter Tarabek
2021 IEEE Access  
Moreover, pre-trained networks are suboptimal and the performance can be considerably improved by training the backbone network for visual tracking from scratch [17] . He et al.  ...  Thus, to exploit deep neural networks for feature extraction that involve padding, translation has to be a part of data augmentation.  ...  The supplementary information about backbone and training serves for more granular comparison and emphasizes existing design trends.  ... 
doi:10.1109/access.2021.3101988 fatcat:iwjlqirwqrav5nadbaw2g5huuu

Visual tracking based on transfer learning of deep salience information

Zuo Haorui, Xu Zhiyong, Zhang Jianlin, Jia Ge
2020 Opto-Electronic Advances  
In this paper, we propose a new visual tracking method in light of salience information and deep learning. Salience detection is used to exploit features with salient information of the image.  ...  Complicated representations of image features can be gained by the function of every layer in convolution neural network (CNN).  ...  Acknowledgements This work was supported by the West Light Foundation for Innovative Talents of the Chinese Academy of Sciences (CAS) (No.YA18K001).  ... 
doi:10.29026/oea.2020.190018 doaj:e230fac26e2948d28ca973523f087436 fatcat:skxrt4q2yrdeheanqjpwfyjnau

Target-Aware Deep Tracking [article]

Xin Li, Chao Ma, Baoyuan Wu, Zhenyu He, Ming-Hsuan Yang
2019 arXiv   pre-print
Existing deep trackers mainly use convolutional neural networks pre-trained for generic object recognition task for representations.  ...  Despite demonstrated successes for numerous vision tasks, the contributions of using pre-trained deep features for visual tracking are not as significant as that for object recognition.  ...  Despite the state-of-the-art performance of existing deep trackers, we note that the contributions of pre-trained deep features for visual tracking are not as significant as that for object recognition  ... 
arXiv:1904.01772v1 fatcat:ttqdj7sc4jaxxanbaamsuiv6ye
« Previous Showing results 1 — 15 out of 6,492 results