影像物件追蹤從以前到現在一直是被研究的領域，從原本利用傳統影像處理的方法、評估判別模型到現在的深度學習。而在追蹤的期間，時常會遇到追蹤的物件被障礙物遮擋的問題，大多論文中處理遮蔽物的方法是評估目前目標物的位置或者搜尋整張影像找尋最類似目標物。假如遇到的遮蔽物與目標本身類似，這會導致追蹤演算法無法認清目標物跟遮蔽物的差別。而如果將此資訊再重新餵給追蹤模型，最終追蹤模型可能會因為訓練樣本而造成追蹤目標的準確率下降。如果我們能夠更精準的考慮何時目標物被遮蔽物遮擋，是否能更將模型訓練好，讓模型得以更加精準的判斷目標物。
深度信息可以提供給我們目標物的3D位置，這給了我們更多的信息和對移動方向的更好的理解。在本文中，我們提出了一種基於深度及卡爾曼的遮擋來處理遮蔽物問題的視覺追蹤算法來幫助我們追蹤目標物。追蹤的模型是基於判別式模型來進行追蹤，遮蔽物處理的部分，除了顏色訊息以外，利用深度訊息來定義目標物何時被遮蔽，並且利用卡爾曼來評估被遮蔽的目標物的路線。除了卡爾曼預測的位置以外，還有多方位的位置評估來更加的精準找到目標物。實驗數據最終顯示我們提出的方法能夠提升追蹤目標物的效果。

Visual tracking has been researched over through years in which one of the most challenging problems is occlusion handling. A tracking target in a complex scene would be occluded by environmental objects such as buildings, people, etc, which often causes the tracking algorithms to lose the track of the target. Different kinds of occlusion handling techniques have been proposed. However, most of these methods are vision-based that search the whole scene and compare the similarities between the target and the potential candidate bounding box. These methods often ignored the depth information of the target and environment objects. The depth information can provide a 3D position of the target, which gives us more information and a better understanding of the moving direction of the target. In this thesis, a visual tracking algorithm with depth-based occlusion handling ability to help us track targets has been proposed.
Our tracker is based on a discriminative model of a siamese network architecture to detect the target position which takes into consideration of both color and depth information of objects in an image. In the occlusion handling part, the trajectory estimation is based on the Kalman filter. We proposed a multi-position handling method that can enhance the estimation accuracy of the search for the target if it is occluded and becomes losing track.
The final results show that the average improvement on the F1-score is 5.2 percent against a baseline model Dephtrack Tracker (DeT) over the benchmark depthtrack dataset, therefore our proposed method can improve the performance of the RGB-D visual object tracking.

摘要
Abstract
Acknowledgement
List of Tables
List of Figures
1 Introduction--------------- 1
2 Related Work--------------- 5
3 Methodology---------------- 11
4 Experiments and Results---- 34
5 Conclusion and Future Work- 43
References------------------- 46

[1] Yang Li and Jianke Zhu. A scale adaptive kernel correlation filter tracker with feature
integration. In European conference on computer vision, pages 254–265. Springer,
2014.
[2] Hamed Kiani Galoogahi, Ashton Fagg, and Simon Lucey. Learning background-
aware correlation filters for visual tracking. In Proceedings of the IEEE international
conference on computer vision, pages 1135–1143, 2017.
[3] Shuran Song and Jianxiong Xiao. Tracking revisited using rgbd camera: Unified
benchmark and baselines. In Proceedings of the IEEE international conference on
computer vision, pages 233–240, 2013.
[4] Alan Lukezic, Ugur Kart, Jani Kapyla, Ahmed Durmush, Joni-Kristian Kamarainen,
Jiri Matas, and Matej Kristan. Cdtb: A color and depth visual object tracking dataset
and benchmark. In Proceedings of the IEEE/CVF International Conference on Com-
puter Vision, pages 10013–10022, 2019.
[5] Song Yan, Jinyu Yang, Jani K ̈apyl ̈a, Feng Zheng, Ales Leonardis, and Joni-Kristian
K ̈am ̈ar ̈ainen. Depthtrack : Unveiling the power of RGBD tracking. CoRR,
abs/2108.13962, 2021. URL https://arxiv.org/abs/2108.13962.
[6] Goutam Bhat, Martin Danelljan, Luc Van Gool, and Radu Timofte. Learning dis-
criminative model prediction for tracking. CoRR, abs/1904.07220, 2019. URL
http://arxiv.org/abs/1904.07220.
[7] Martin Danelljan, Goutam Bhat, Fahad Shahbaz Khan, and Michael Felsberg. ATOM:
accurate tracking by overlap maximization. CoRR, abs/1811.07628, 2018. URL
http://arxiv.org/abs/1811.07628.
[8] Greg Welch, Gary Bishop, et al. An introduction to the kalman filter. 1995.
[9] Zhaoxia Fu and Yan Han. Centroid weighted kalman filter for visual object track-
ing. Measurement, 45(4):650–655, 2012. ISSN 0263-2241. doi: https://doi.org/
10.1016/j.measurement.2012.01.004. URL https://www.sciencedirect.
com/science/article/pii/S026322411200005X.
[10] Zebin Cai, Zhenghui Gu, Zhu Yu, Hao Liu, and Ke Zhang. A real-time visual object
tracking system based on kalman filter and mb-lbp feature matching. Multimedia
Tools and Applications, 75, 12 2014. doi: 10.1007/s11042-014-2411-6.
[11] Dorin Comaniciu and Visvanathan Ramesh. Mean shift and optimal prediction for
efficient object tracking. In Proceedings 2000 International Conference on Image
Processing (Cat. No. 00CH37101), volume 3, pages 70–73. IEEE, 2000.
[12] Huiyu Zhou, Yuan Yuan, and Chunmei Shi. Object tracking using sift features and
mean shift. Computer vision and image understanding, 113(3):345–352, 2009.
[13] Changjiang Yang, R. Duraiswami, and L. Davis. Fast multiple object tracking via
a hierarchical particle filter. In Tenth IEEE International Conference on Computer
Vision (ICCV’05) Volume 1, volume 1, pages 212–219 Vol. 1, 2005. doi: 10.1109/
ICCV.2005.95.
[14] Zulfiqar Hasan Khan, Irene Yu-Hua Gu, and Andrew G. Backhouse. Robust visual
object tracking using multi-mode anisotropic mean shift and particle filters. IEEE
Transactions on Circuits and Systems for Video Technology, 21(1):74–87, 2011. doi:
10.1109/TCSVT.2011.2106253.
[15] Irene Anindaputri Iswanto and Bin Li. Visual object tracking based on mean-shift and
particle-kalman filter. Procedia computer science, 116:587–595, 2017.
[16] Martin Danelljan, Gustav Hager, Fahad Shahbaz Khan, and Michael Felsberg. Con-
volutional features for correlation filter based visual tracking. In Proceedings of the
IEEE International Conference on Computer Vision (ICCV) Workshops, December
2015.
[17] Alan Lukezic, Tomas Vojir, Luka ˇCehovin Zajc, Jiri Matas, and Matej Kristan.
Discriminative correlation filter with channel and spatial reliability. In Proceedings of
the IEEE conference on computer vision and pattern recognition, pages 6309–6318,
2017.
[18] Matthias Mueller, Neil Smith, and Bernard Ghanem. Context-aware correlation filter
tracking. In Proceedings of the IEEE conference on computer vision and pattern
recognition, pages 1396–1404, 2017.
[19] Luca Bertinetto, Jack Valmadre, Joao F Henriques, Andrea Vedaldi, and Philip HS
Torr. Fully-convolutional siamese networks for object tracking. In European confer-
ence on computer vision, pages 850–865. Springer, 2016.
[20] Qing Guo, Wei Feng, Ce Zhou, Rui Huang, Liang Wan, and Song Wang. Learning
dynamic siamese network for visual object tracking. In Proceedings of the IEEE
international conference on computer vision, pages 1763–1771, 2017.
[21] Anfeng He, Chong Luo, Xinmei Tian, and Wenjun Zeng. A twofold siamese network
for real-time object tracking. In Proceedings of the IEEE Conference on Computer
Vision and Pattern Recognition (CVPR), June 2018.
[22] Xingping Dong and Jianbing Shen. Triplet loss in siamese network for object track-
ing. In Proceedings of the European conference on computer vision (ECCV), pages
459–474, 2018.
[23] Bin Yan, Houwen Peng, Jianlong Fu, Dong Wang, and Huchuan Lu. Learning spatio-
temporal transformer for visual tracking. In Proceedings of the IEEE/CVF Interna-
tional Conference on Computer Vision, pages 10448–10457, 2021.
[24] Feng Xiao, Qiuxia Wu, and Han Huang. Single-scale siamese network based rgb-d
object tracking with adaptive bounding boxes. Neurocomputing, 451:192–204, 2021.
[25] Luca Bertinetto, Jack Valmadre, Joo F. Henriques, Andrea Vedaldi, and Philip H. S.
Torr. Fully-convolutional siamese networks for object tracking, 2016. URL https:
//arxiv.org/abs/1606.09549.
[26] Borui Jiang, Ruixuan Luo, Jiayuan Mao, Tete Xiao, and Yuning Jiang. Acquisition of
localization confidence for accurate object detection. In Proceedings of the European
conference on computer vision (ECCV), pages 784–799, 2018.
[27] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning
for image recognition. In Proceedings of the IEEE conference on computer vision
and pattern recognition, pages 770–778, 2016.
[28] Borui Jiang, Ruixuan Luo, Jiayuan Mao, Tete Xiao, and Yuning Jiang. Acquisition of
localization confidence for accurate object detection. In Proceedings of the European
conference on computer vision (ECCV), pages 784–799, 2018.
[29] Matej Kristan, Jiri Matas, Aleˇs Leonardis, Tomas Vojir, Roman Pflugfelder, Gustavo
Fernandez, Georg Nebehay, Fatih Porikli, and Luka ˇCehovin. A novel performance
evaluation methodology for single-target trackers. IEEE Transactions on Pattern
Analysis and Machine Intelligence, 38(11):2137–2155, Nov 2016. ISSN 0162-8828.
doi: 10.1109/TPAMI.2016.2516982.