-
摘要: 边界框回归分支是深度目标跟踪器的关键模块, 其性能直接影响跟踪器的精度. 评价精度的指标之一是交并比(Intersection over Union, IoU). 基于 IoU 的损失函数取代了
$ \ell_n $ -norm 损失成为目前主流的边界框回归损失函数, 然而 IoU 损失函数存在两个固有缺陷: 一个是当预测框与真值框不相交时 IoU 为常量 0, 无法梯度下降更新边界框的参数; 另一个是在 IoU 取得最优值时其梯度不存在, 边界框很难收敛到 IoU 最优处. 本文揭示了在回归过程中 IoU 最优的边界框各参数之间蕴含的定量关系, 指出在边界框中心处于特定位置时存在多种尺寸不同的边界框使 IoU 损失最优的情况, 这增加了边界框尺寸回归的不确定性. 本文从优化两个统计分布之间散度的视角看待边界框回归问题, 提出了光滑 IoU 损失, 即构造了在全局上光滑 (即连续可微) 且极值唯一的损失函数, 该损失函数自然蕴含边界框各参数之间特定的最优关系, 其唯一取极值的边界框可使 IoU 达到最优. 光滑性确保了在全局上梯度存在使得边界框更容易回归到极值处, 而极值唯一确保了在全局上可梯度下降更新参数, 从而避开了 IoU 损失的固有缺陷. 提出的光滑 IoU 损失可以很容易取代 IoU 损失集成到现有的深度目标跟踪器上训练边界框回归, 在 LaSOT, GOT-10k, TrackingNet 和 OTB2015 等测试基准上所取得的结果验证了光滑 IoU 损失的易用性和有效性.-
关键词:
- 光滑IoU 损失 /
- $ \ell_n $-norm 损失 /
- 边界框回归 /
- 目标跟踪
Abstract: The branch of bounding box regression is a critical module in visual object trackers, and its performance directly affects accuracy of a tracker. One of evaluation metrics used to measure accuracy is Intersection over Union (IoU). The IoU loss which was proposed to replace$ \ell_n $ -norm loss for bounding box regression is increasingly popular. However, there are two inherent issues in IoU loss: one is that the parameters of bounding box can not be updated via gradient descent if the predicted box does not intersect with ground-truth box; the other is the gradient of the optimal IoU does not exist, so that it is difficult to make the predicted box regressed to the IoU optimum. We reveal the explicit relationship among the parameters of IoU optimal bounding box in regression process, and point out that the size of a predicted box which make IoU loss optimal is not unique when its center is in specific areas, increasing the uncertainty of bounding box regression. From the perspective of optimizing divergence between two distributions, we propose a Smooth-IoU loss, which is a globally smooth (i.e., continuously differentiable) loss function with unique extremum. The Smooth-IoU loss naturally implicates a specific optimal relationship among the parameters of bounding box, and its gradient over the global domain exists, making it easier to regress the predicted box to the extremal bounding box, and the unique extremum ensures that the parameters can be updated via gradient descent. In addition, the proposed smooth-IoU loss can be easily incorporated into existing trackers by replacing the IoU-based loss to train bounding box regression. Extensive experiments on visual tracking benchmarks including OTB2015, TrackingNet, GOT-10k, and LaSOT demonstrate that Smooth-IoU loss achieves state-of-the-art performance, confirming its effectiveness and efficiency.-
Key words:
- Smooth-IoU loss /
- $ \ell_n $-norm loss /
- bounding box regression /
- visual tracking
1) 收稿日期 2021-06-11 录用日期 2021-09-17 Manuscript received June 11, 2021; accepted September 17, 2021 国家自然科学基金重点项目 (51935005), 基础科研项目(JCKY20200603C010), 空间智能控制技术重点实验室基金 (ZDSYS-2018-02) 资助 Partially supported by the National Natural Science Foundation of China under Grants (No. 51935005), the Basic Scientific Research Projects (No. JCKY20200603C010), and the Mutual2) Fund of Space Intelligent Control Technology Key Laboratory (No. ZDSYS-2018-02) 本文责任编委 桑农 Recommended by Associate Editor SANG Nong 1. 哈尔滨工业大学计算学部模式识别与智能系统研究中心 哈尔滨 150001 1. Pattern Recognition and Intelligence System Research Center, Faculty of Computing, Harbin Institute of Technology, Harbin 1500013) 1 对于不同的研究对象, 光滑的含义也有所区别, 在本文中称在开集$ X\in \mathbb{R}^n $ 上的函数$ f:X\to\mathbb{R} $ 是光滑的, 如果$ f $ 是$ C^1 $ 类的.$ C^1 $ 类的函数必然是可微的. 在本文的定义下光滑性也可以称作连续可微性 -
图 9 比较各种边界框回归损失i.e.
$ {\cal{L}}_{{\rm{IoU}}} $ ,$ {\cal{L}}_{{\rm{GIoU}}} $ ,$ {\cal{L}}_{{\rm{DIoU}}} $ ,$ {\cal{L}}_{{\rm{CIoU}}} $ 和提出的$ {\cal{L}}_{{\rm{SIoU}}} $ )的收敛能力Fig. 9 Comparison among the convergence performance of different bounding box regression losses(i.e.
$ {\cal{L}}_{{\rm{IoU}}} $ ,$ {\cal{L}}_{{\rm{GIoU}}} $ ,$ {\cal{L}}_{{\rm{DIoU}}} $ ,$ {\cal{L}}_{{\rm{CIoU}}} $ , and proposed$ {\cal{L}}_{{\rm{SIoU}}} $ )表 1 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamFC++在基准 LaSOT 上的测试结果Table 1 Comparison between the performance of SiamFC++ trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test set of LaSOTLoss $ \backslash $ Metric $ Succ. $ $ Prec. $ $ P_{norm} $ ${\cal{L} }_{{\rm{IoU}}}$ 55.6 55.5 64.8 ${\cal{L} }_{{\rm{SIoU}}}$ 57.6 58.3 66.9 Rel. Improv. (%) 3.60% 5.05% 3.24% 表 2 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamBAN 在基准 LaSOT 上的测试结果Table 2 Comparison between the performance of SiamBAN trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test set of LaSOTLoss $ \backslash $ Metric $ Succ. $ $ Prec. $ $ P_{norm} $ ${\cal{L} }_{{\rm{IoU}}}$ 51.4 52.1 59.8 ${\cal{L} }_{{\rm{SIoU}}}$ 54.3 53.9 63.3 Rel. Improv. (%) 5.64% 3.45% 4.85% 表 3 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamCAR 在基准 LaSOT 上的测试结果Table 3 Comparison between the performance of SiamCAR trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test set of LaSOTLoss $ \backslash $ Metric $ Succ. $ $ Prec. $ $ P_{norm} $ ${\cal{L} }_{{\rm{IoU}}}$ 51.6 52.4 61.0 ${\cal{L} }_{{\rm{SIoU}}}$ 54.9 54.8 63.1 Rel. Improv. (%) 6.39 % 4.58% 3.44 % 表 4 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamFC++ 在 GOT-10k 上测试结果Table 4 Comparison between the performance of SiamFC++ trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test set of GOT-10kLoss $ \backslash $ Metric $ {\rm{AO}} $ $ {\rm{SR}}_{.50} $ $ {\rm{SR}}_{.75} $ ${\cal{L} }_{{\rm{IoU}}}$ 59.5 69.5 47.9 ${\cal{L} }_{{\rm{SIoU}}}$ 61.7 74.7 46.8 Rel. Improv. (%) 3.69% 7.48% −2.29% 表 5 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamCAR 在 GOT-10k 上测试结果Table 5 Comparison between the performance of SiamCAR trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test set of GOT-10kLoss $ \backslash $ Metric $ {\rm{AO}} $ $ {\rm{SR}}_{.50} $ $ {\rm{SR}}_{.75} $ ${\cal{L} }_{{\rm{IoU}}}$ 58.1 68.3 44.1 ${\cal{L} }_{{\rm{SIoU}}}$ 60.2 72.6 46.4 Rel. Improv. (%) 3.61% 6.29% 5.22% 表 6 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamFC++ 在TrackingNet上测试结果Table 6 Comparison between the performance of SiamFC++ trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on the test of TrackingNet.Loss$ \backslash $Metric $ Prec. $ $ P_{norm} $ $ Succ. $ ${\cal{L} }_{\rm{{IoU}}}$ 70.5 80.0 75.4 ${\cal{L} }_{{\rm{SIoU}}}$ 72.1 81.9 76.2 Rel. Improv. (%) 2.27% 2.37% 1.06% 表 7 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamFC++ 在OTB2015上测试结果Table 7 Comparison between the performance of SiamFC++ trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on OTB2015.Loss $ \backslash $ Metric $ Succ. $ $ Prec. $ ${\cal{L} }_{{\rm{IoU}}}$ 68.2 89.5 ${\cal{L} }_{{\rm{SIoU}}}$ 68.7 89.8 Rel. Improv. (%) 0.74% 0.34% 表 8 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamBAN 在OTB2015上测试结果Table 8 Comparison between the performance of SiamBAN trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on OTB2015.Loss $ \backslash $ Metric $ Succ. $ $ Prec. $ ${\cal{L} }_{{\rm{IoU}}}$ 69.6 91.0 ${\cal{L} }_{{\rm{SIoU}}}$ 69.9 91.5 Rel. Improv. (%) 0.43% 0.55% 表 9 与先进方法在基准 LaSOT 上的性能评估对比
Table 9 Performance evaluation for state-of-the-art algorithms on LaSOT. The best three results are marked in bold, italics and underline fonts, respectively
Metric$ \backslash $Method SiamBAN ATOM SiamCAR SiamRPN++ Ocean-online SiamFC++ DiMP SiamBAN
(SIoU)SiamCAR
(SIoU)SiamFC++
(SIoU)$ Succ. $ 51.4 51.5 51.6 49.6 56.0 55.6 56.8 54.3 54.2 57.6 $ Prec. $ 52.1 50.5 52.4 49.1 56.6 55.5 56.4 53.9 53.7 58.3 $ P_{norm} $ 59.8 57.6 61.0 56.9 65.1 64.8 64.3 63.3 63.1 66.9 表 10 与先进方法在基准 GOT-10k 上的性能评估对比
Table 10 Performance evaluation for state-of-the-art algorithms on GOT-10k. The best three results are marked in bold, italics and underline fonts, respectively
Metric$ \backslash $Method MDNet SPM ATOM SiamCAR SiamRPN++ Ocean-online D3S SiamFC++ DiMP-50 SiamCAR
(SIoU)SiamFC++
(SIoU)$ {\rm{AO}} $ 29.9 51.3 55.6 56.9 51.7 61.1 59.7 59.5 61.1 60.2 61.7 $ {\rm{SR}}_{.50} $ 30.3 59.3 63.4 67.0 61.8 72.1 67.6 69.5 71.2 72.6 74.7 表 11 与先进方法在基准 TrackingNet 上的性能评估对比
Table 11 Performance evaluation for state-of-the-art algorithms on TrackingNet. The best three results are marked in bold, italics and underline fonts, respectively
Metric$ \backslash $Method MDNet ATOM DaSiamRPN SiamRPN++ UpdateNet SPM SiamFC++ DiMP SiamFC++(SIoU) $ Succ. $ 60.6 70.3 63.8 73.3 67.7 71.2 75.4 74.0 76.2 $ P_{norm} $ 70.5 77.1 73.3 80.0 75.2 77.8 80.0 80.1 81.9 表 12 分别以
$ {\cal{L}}_{{\rm{IoU}}} $ 和$ {\cal{L}}_{{\rm{SIoU}}} $ 训练的模型 SiamFC++ 在VOT2018上测试结果Table 12 Comparison between the performance of SiamFC++ trained using
$ {\cal{L}}_{{\rm{IoU}}} $ (original),$ {\cal{L}}_{{\rm{SIoU}}} $ on VOT2018Loss$ \backslash $Metric $ \mathbf{A}\uparrow $ $ \mathbf{R}\downarrow $ $ \mathbf{EAO}\uparrow $ ${\cal{L} }_{{\rm{IoU}}}$ 0.586 0.201 0.427 ${\cal{L} }_{{\rm{SIoU}}}$ 0.582 0.196 0.400 表 13 与其它基于 IoU 损失在基准 LaSOT 上训练得到的满足不同 IoU 阈值的测试集图像帧数占比的对比结果
Table 13 Comparison results with other IOU-based loss for the ratio of frames exceeding different IOU thresholds on the test set of LaSOT. The best result is marked in bold
Ratio of Method (%) $ \backslash $IoU ≥ 0.95 ≥ 0.90 ≥ 0.85 ≥ 0.80 ≥ 0.75 ≥ 0.70 ≥ 0.65 ≥ 0.60 ≥ 0.55 ≥ 0.50 SiamFC++(SIoU) 2.75 15.93 31.83 44.05 52.33 57.71 61.71 64.41 66.52 68.14 SiamFC++(DIoU) 1.60 13.19 29.72 42.84 51.48 57.09 61.28 64.17 66.31 67.95 SiamFC++(GIoU) 2.45 16.18 31.10 42.26 50.39 55.81 59.56 62.37 64.58 66.34 SiamFC++ 1.52 12.73 29.10 41.90 50.20 55.63 59.72 62.51 64.68 66.31 SiamBAN(SIoU) 1.18 10.79 24.86 36.64 45.50 51.87 56.45 60.03 62.77 64.84 SiamBAN(GIoU) 1.49 11.77 24.71 35.15 44.77 50.79 54.93 57.76 60.70 63.98 SiamBAN 1.98 12.89 25.40 35.57 43.46 49.29 53.38 56.53 58.92 60.78 SiamCAR(SIoU) 1.20 10.80 24.81 36.74 45.75 52.29 56.97 60.71 63.53 65.66 SiamCAR(DIoU) 1.20 10.91 25.10 36.62 45.04 51.47 56.18 59.89 62.70 64.83 SiamCAR 1.27 10.62 23.90 35.98 44.93 50.87 55.08 57.86 59.94 61.61 表 14 与其它基于 IoU 损失在基准 GOT-10k 上训练得到的满足不同 IoU 阈值的测试集图像帧数占比的对比结果
Table 14 Comparison results with other IOU-based loss for the ratio of frames exceeding different IOU thresholds on the test set of GOT-10k. The best result is marked in bold
Ratio of Method (%) $ \backslash $IoU ≥ 0.95 ≥ 0.90 ≥ 0.85 ≥ 0.80 ≥ 0.75 ≥ 0.70 ≥ 0.65 ≥ 0.60 ≥ 0.55 ≥ 0.50 SiamFC++(SIoU) 0.94 8.18 22.01 35.76 46.83 55.71 62.16 67.40 71.39 74.68 SiamFC++(DIoU) 0.72 7.56 21.10 35.11 46.86 55.45 61.68 66.82 70.74 73.86 SiamFC++(GIoU) 0.97 7.20 21.80 34.19 45.85 54.50 59.24 63.48 66.02 69.49 SiamCAR(SIoU) 0.94 8.58 22.20 35.83 46.46 54.74 60.66 65.49 69.30 72.62 SiamCAR(GIoU) 1.13 6.72 19.23 34.78 45.37 53.73 59.98 64.96 68.95 71.96 SiamCAR(DIoU) 0.92 6.19 18.85 32.54 43.73 52.51 58.86 64.04 68.09 71.33 SiamCAR 0.81 8.02 20.76 33.88 44.07 51.87 57.21 61.35 64.99 68.31 表 15 在GOT-10k上对
${\cal{L}}_{{\rm{SIoU}}}$ 的正则项和代理函数消融实验Table 15 Ablation studies about the regulariztion and surrogate function on GOT-10k
Loss $ \backslash $ Metric $ {\rm{AO}} $ $ {\rm{SR}}_{.50} $ $ {\rm{SR}}_{.75} $ $ {\cal{L}}_{{\rm{SIoU}}}({\rm{w/o}} \, {\cal{AR}}) $ 59.9 72.1 46.3 $ {\cal{L}}_{{\rm{SIoU}}} ({\rm{w/}} \;{\cal{AR}}) $ 61.7 74.7 46.8 Rel. Improv. (%) 3.01% 3.61% 1.08% $ {\cal{L}}_{{\rm{SIoU}}}({\rm{w/}} \;{\cal{R}}) $ 61.4 74.5 46.7 Rel. Improv. (%) 2.51% 3.33% 0.86% $ {\cal{L}}_{{\rm{SIoU}}} ({\rm{w/}} \;{\cal{A}}_2) $ 60.3 72.6 46.5 Rel. Improv. (%) 0.67% 0.69% 0.43% $ {\cal{L}}_{{\rm{SIoU}}} ({\rm{w/}} \;{\cal{A}}_4) $ 60.6 73.4 46.3 Rel. Improv. (%) 1.17 % 1.80% 0.0 % $ {\cal{L}}_{{\rm{SIoU}}} ({\rm{w/}} \;{\cal{A}}_8) $ 60.4 73.4 46.0 Rel. Improv. (%) 0.83 % 1.80 % −0.65 % $ {\cal{L}}_{{\rm{SIoU}}} ({\rm{w/}} \;{\cal{A}}_1) $ 58.9 71.3 43.7 Rel. Improv. (%) −1.17 % −1.11 % −5.62% -
[1] 孟琭, 杨旭. 目标跟踪算法综述. 自动化学报, 2019, 45(7):1244-1260Meng Lu, Yang Xu. A survey of object tracking algorithms. Acta Automatica Sinica, 2019, 45(7):1244-1260 [2] 蒋弘毅, 王永娟, 康锦煜. 目标检测模型及其优化方法综述. 自动化学报, 2021, 47(6): 1232-1255Jiang Hong-Yi, Wang Yong-Juan, Kang Jin-Yu. A survey of object detection models and its optimization methods. Acta Automatica Sinica, 2021, 47(6):1232-1255 [3] Girshick R. B. Fast R-CNN. In: Proceedings of 2015 IEEE International Conference on Computer Vision. Santiago, Chile: IEEE Computer Society, 2015. 1440−1448 [4] Yu J, Jiang Y, Wang Z, et al. Unitbox: An advanced object detection network. In: Proceedings of the 2016 ACM Conference on Multimedia Conference. Amsterdam, The Netherlands: ACM, 2016. 516−520 [5] Rezatofighi H, Tsoi N, Gwak J, et al. Generalized intersection over union: A metric and a loss for bounding box regression. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 658−666 [6] Zheng Z, Wang P, Liu W, et al. Distance-iou loss: Faster and better learning for bounding box regression. In: Proceedings of The Thirty-Fourth AAAI Conference on Artificial Intelligence. New York, NY, USA: AAAI Press, 2020. 12993−13000 [7] Zhang Y, Ren W, Zhang Z, et al. Focal and efficient IOU loss for accurate bounding box regression[Online], available: https://arxiv.org/abs/2101.08158, August 11, 2021 [8] Li B, Yan J, Wu W, et al. High performance visual tracking with siamese region proposal network. In: Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE Computer Society, 2018. 8971−8980 [9] Li B, Wu W, Wang Q, et al. Siamrpn++: Evolution of siamese visual tracking with very deep networks. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition.Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 4282−4291 [10] Zhu Z, Wang Q, Li B, et al. Distractor-aware siamese networks for visual object tracking. In: Proceedings of 2018 - 15th European Conference on Computer Vision, Part IX. Munich, Germany: Springer, 2018. 103−119 [11] He Y, Zhu C, Wang J, et al. Bounding box regression with uncertainty for accurate object detection. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 2888−2897 [12] Law H, Deng J. Cornernet: Detecting objects as paired keypoints. In: Proceedings of 2018 - 15th European Conference on Computer Vision, Part XIV. Munich, Germany: Springer, 2018. 765−781 [13] Gidaris S, Komodakis N. Locnet: Improving localization accuracy for object detection. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE Computer Society, 2016. 789−798 [14] Zhou X, Koltun V, Krähenbühl P. Tracking objects as points. In: Proceedings of 2020 - 16th European Conference on Computer Vision, Part IV. Glasgow, UK: Springer, 2020. 474−490 [15] Lin T.-Y, Goyal P, Girshick R, et al. Focal loss for dense object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(2):318-327 doi: 10.1109/TPAMI.2018.2858826 [16] Held D, Thrun S, Savarese S. Learning to track at 100 FPS with deep regression networks. In: Proceedings of 2016 - 14th European Conference on Computer Vision, Part I. Amsterdam, The Netherlands: Springer, 2016. 749−765 [17] Bertinetto L, Valmadre J, Henriques J. F, et al. Fully-convolutional siamese networks for object tracking. In: Proceedings of 2016 European Conference on Computer Vision Workshops, Part Ⅱ. Amsterdam, The Netherlands: Springer, 2016. 850−865 [18] Jiang B, Luo R, Mao J, et al. Acquisition of localization confidence for accurate object detection. In: Proceedings of 2018 - 15th European Conference on Computer Vision, Part XIV. Munich, Germany: Springer, 2018. 816−832 [19] Wang G, Luo C, Xiong Z, et al. Spm-tracker: Series-parallel matching for real-time visual object tracking. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 3643−3652 [20] Xu Y, Wang Z, Li Z, et al. Siamfc++: Towards robust and accurate visual tracking with target estimation guidelines. In: Proceedings of The Thirty-Fourth AAAI Conference on Artificial Intelligence. New York, NY, USA: AAAI Press, 2020. 12549−12556 [21] Chen Z, Zhong B, Li G, et al. Siamese box adaptive network for visual tracking. In: Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE, 2020. 6667−6676 [22] Guo D, Wang J, Cui Y, et al. Siamcar: Siamese fully convolutional classification and regression for visual tracking. In: Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE, 2020. 6268−6276 [23] Zhang Z, Peng H, Fu J, et al. Ocean: Object-aware anchor-free tracking. In: Proceedings of 2020 - 16th European Conference on Computer Vision, Part XXI. Glasgow, UK: Springer, 2020. 771−787 [24] 谭建豪, 郑英帅, 王耀南, 马小萍. 基于中心点搜索的无锚框全卷积孪生跟踪器. 自动化学报, 2021, 47(4): 801-812Tan Jian-Hao, Zheng Ying-Shuai, Wang Yao-Nan, et al. AFST: Anchor-free fully convolutional siamese tracker with searching center point. Acta Automatica Sinica, 2021, 47(4):801-812 [25] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions. In: Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA: IEEE Computer Society, 2015. 1−9 [26] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE Computer Society, 2016. 770−778 [27] Russakovsky O, Deng J, Su H, et al. Imagenet large scale visual recognition challenge. International Journal of Computer Vision, 2015, 115(3):211-252 doi: 10.1007/s11263-015-0816-y [28] Lin T, Maire M, Belongie S. J, et al. Microsoft COCO: common objects in context. In: Proceedings of 2014 - 13th European Conference on Computer Vision, Part V. Zurich, Switzerland: Springer, 2014. 740−755 [29] Real E, Shlens J, Mazzocchi S, et al. Youtube-boundingboxes: A large high-precision human-annotated data set for object detection in video. In: Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE Computer Society, 2017. 7464−7473 [30] Fan H, Lin L, Yang F, et al. Lasot: A high-quality benchmark for large-scale single object tracking. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 5374−5383 [31] Müller M, Bibi A, Giancola S, et al. Trackingnet: A large-scale dataset and benchmark for object tracking in the wild. In: Proceedings of 2018 - 15th European Conference on Computer Vision, Part I. Munich, Germany: Springer, 2018. 310−327 [32] Huang L, Zhao X, Huang K. Got-10k: A large high-diversity benchmark for generic object tracking in the wild. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(5):1562-1577 doi: 10.1109/TPAMI.2019.2957464 [33] Wu Y, Lim J, Yang M. Object tracking benchmark. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1834-1848 doi: 10.1109/TPAMI.2014.2388226 [34] Kristan M, He Z. The sixth visual object tracking VOT2018 challenge results. In: Proceedings of 2018 European Conference on Computer Vision Workshops, Part I. Munich, Germany: Springer, 2018. 3−53 [35] Wang Q, Zhang L, Bertinetto L, et al. Fast online object tracking and segmentation: A unifying approach. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 1328−1338 [36] Huang L, Zhao X, Huang K. Globaltrack: A simple and strong baseline for long-term tracking. In: Proceedings of The Thirty-Fourth AAAI Conference on Artificial Intelligence. New York, NY, USA: AAAI Press, 2020. 11037−11044 [37] Fan H, Ling H. Siamese cascaded region proposal networks for real-time visual tracking. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 7952−7961 [38] Danelljan M, Bhat G, Khan F. S, et al. ATOM: accurate tracking by overlap maximization. In: Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, USA: Computer Vision Foundation/IEEE, 2019. 4660−4669 [39] Bhat G, Danelljan M, Gool L. V, et al. Learning discriminative model prediction for tracking. In: Proceedings of 2019 IEEE/CVF International Conference on Computer Vision. Seoul, Korea (South): IEEE, 2019. 6181−6190 [40] Nam H, Han B. Learning multi-domain convolutional neural networks for visual tracking. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE Computer Society, 2016. 4293−4302 [41] Lukezic A, Matas J, Kristan M. D3S - A discriminative single shot segmentation tracker. In: Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE, 2020. 7131−7140 [42] Zhang L, Gonzalez -Garcia A, Weijervan de J, et al. Learning the model update for siamese trackers. In: Proceedings of 2019 IEEE/CVF International Conference on Computer Vision. Seoul, Korea (South): IEEE, 2019. 4009−4018 -
AAS-CN-2021-0525数据.zip
-

计量
- 文章访问数: 377
- HTML全文浏览量: 158
- 被引次数: 0