Deep Meta Learning for Real-Time Target-Aware Visual Tracking

Choi, Janghoon; Kwon, Junseok; Lee, Kyoung Mu

doi:10.1109/iccv.2019.00100

Cited by 117 publications

(73 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SCSAtt also achieved 7.31%, 5.99%, 11.69% and 7.31% progress in success score and 10.55%, 4.68%, 13.11% and 6.84% progress in precision score than the MemTrack [40], CREST [61], SRDCF [62], and DsiamM [7] trackers, respectively. Moreover, the proposed method has shown that the performance improvement of 5.34% and 2.56%, 4.28% and 3.03%, 10.99% and 8.66%, and 23.21% and 17.58% in precision and success than the most recent trackers including DSAR-CF [63], MemDTC [45], MLT [65], and UDT [64], respectively. SCSAtt, therefore, constantly outperform on both success and precision scoring metric that demonstrates the effectiveness of our tracker in terms of robustness.…”

Section: B Evaluation On Otb50 Benchmarkmentioning

confidence: 89%

“…We also compared our proposed tracker with the most recent trackers including DSAR-CF [63], MLT [65], and UDT [64]. The proposed tracker achieved 2.76%, 7.28%, and 12.5% improvement in precision score and 0.31%, 6.66%, and 9.20% improvement in success score compared to DSAR-CF, MLT, and UDT trackers, respectively.…”

Section: A Evaluation On Otb100 Benchmarkmentioning

confidence: 98%

“…Success plots show the overall percentage of the overlap score, whereas precision plots show the percentage of center error distance between ground-truth and predicted bounding box. To keep our comparison fair, we accumulated various trackers types, including Siamese based trackers (SiamFC [5], SiamTri [23], SIAMRPN [59], and CFNet [39]), attentional Siamese trackers (MemTrack [40] and MemDTC [45]), correlation filter based trackers (STAPLE [60], CREST [61], SRDCF [62], DSAR-CF [63]), and others (UDT [64], DSiamM [7], and MLT [65]). Fig.…”

Section: A Evaluation On Otb100 Benchmarkmentioning

confidence: 99%

See 2 more Smart Citations

Efficient Visual Tracking With Stacked Channel-Spatial Attention Learning

2020

View full text Add to dashboard Cite

Template based learning, particularly Siamese networks, has recently become popular due to balancing accuracy and speed. However, preserving tracker robustness against challenging scenarios with real-time speed is a primary concern for visual object tracking. Siamese trackers confront difficulties handling target appearance changes continually due to less discrimination ability learning between target and background information. This paper presents stacked channel-spatial attention within Siamese networks to improve tracker robustness without sacrificing fast-tracking speed. The proposed channel attention strengthens target-specific channels increasing their weight while reducing the importance of irrelevant channels with lower weights. Spatial attention is focusing on the most informative region of the target feature map. We integrate the proposed channel and spatial attention modules to enhance tracking performance with end-to-end learning. The proposed tracking framework learns what and where to highlight important target information for efficient tracking. Experimental results on widely used OTB100, OTB50, VOT2016, VOT2017/18, TC-128, and UAV123 benchmarks verified the proposed tracker achieved outstanding performance compared with state-of-the-art trackers. INDEX TERMS Deep learning, Siamese architecture, stacked channel-spatial attention, visual object tracking.

show abstract

Section: B Evaluation On Otb50 Benchmarkmentioning

confidence: 89%

Section: A Evaluation On Otb100 Benchmarkmentioning

confidence: 98%

Section: A Evaluation On Otb100 Benchmarkmentioning

confidence: 99%

See 1 more Smart Citation

Efficient Visual Tracking With Stacked Channel-Spatial Attention Learning

2020

View full text Add to dashboard Cite

show abstract

“…In order to achieve better and more powerful feature embedding, RASNet [44] and DaSiamRPN [56] exploit several types attention mechanisms and more effective training strategy with large-scaled training-sets, respectively. DSiam tracker [14] and MLT tracker [4] update the tracking model via a rapid transformation module and a meta-learner network. The merits of deep networks [25,55] is also expected to be leveraged.…”

Section: Related Work 21 Visual Object Trackingmentioning

confidence: 99%

Reinforced Similarity Learning

Zhang

Zheng

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Recently, Siamese networks based tracking algorithms have shown favorable performance. Latest work focuses on better feature embedding and target state estimation, which greatly improves the accuracy. Nevertheless, the simple cross-correlation operation of the features between a fixed template and the search region limits their robustness and discrimination capability. In this paper, we pay more attention to learn an outstanding similarity measure for robust tracking. We propose a novel relation network that can be integrated on top of previous trackers without any need for further training of the siamese networks, which achieves a superior discriminative ability. During online inference, we utilize the feedback from high-confidence tracking results to obtain an additional template and update it, which improves the robustness and generalization. We implement two versions of the proposed approach with the SiamFC-based tracker and SiamRPN-based tracker to validate the strong compatibility of our algorithm. Extensive experimental results on several tracking benchmarks indicate that the proposed method can effectively improve the performance and robustness of the underlying trackers without reducing speed too much, and performs superiorly against the state-of-the-art trackers.

show abstract

“…Most traditional algorithms for ground-based cloud recognition utilize hand-crafted features, for example, brightness, texture, shape and color, to represent cloud images [12][13][14][15][16][17][18], but they are deficient in modeling complex data distribution. Recently, the convolutional neural network (CNN) [19][20][21][22][23][24] has achieved remarkable performance in various research fields due to the nature of learning highly nonlinear feature transformations. Inspired by this, some cloud-related researchers employ the CNNs to exploit visual features from ground-based cloud images and thrust the performance of ground-based cloud recognition to a new level.…”

Section: Introductionmentioning

confidence: 99%

Multi-Evidence and Multi-Modal Fusion Network for Ground-Based Cloud Recognition

et al. 2020

View full text Add to dashboard Cite

In recent times, deep neural networks have drawn much attention in ground-based cloud recognition. Yet such kind of approaches simply center upon learning global features from visual information, which causes incomplete representations for ground-based clouds. In this paper, we propose a novel method named multi-evidence and multi-modal fusion network (MMFN) for ground-based cloud recognition, which could learn extended cloud information by fusing heterogeneous features in a unified framework. Namely, MMFN exploits multiple pieces of evidence, i.e., global and local visual features, from ground-based cloud images using the main network and the attentive network. In the attentive network, local visual features are extracted from attentive maps which are obtained by refining salient patterns from convolutional activation maps. Meanwhile, the multi-modal network in MMFN learns multi-modal features for ground-based cloud. To fully fuse the multi-modal and multi-evidence visual features, we design two fusion layers in MMFN to incorporate multi-modal features with global and local visual features, respectively. Furthermore, we release the first multi-modal ground-based cloud dataset named MGCD which not only contains the ground-based cloud images but also contains the multi-modal information corresponding to each cloud image. The MMFN is evaluated on MGCD and achieves a classification accuracy of 88.63% comparative to the state-of-the-art methods, which validates its effectiveness for ground-based cloud recognition.

show abstract

Deep Meta Learning for Real-Time Target-Aware Visual Tracking

Cited by 117 publications

References 47 publications

Efficient Visual Tracking With Stacked Channel-Spatial Attention Learning

Efficient Visual Tracking With Stacked Channel-Spatial Attention Learning

Reinforced Similarity Learning

Multi-Evidence and Multi-Modal Fusion Network for Ground-Based Cloud Recognition

Contact Info

Product

Resources

About