Abstract
The realization of human-machine-environment intimate interaction by intelligent robots is the research direction of cutting-edge exploration in the field of robotics. One of the important tasks is to realize active target tracking on the robot platform. Single-target tracking is subject to data changes such as target position and size in the video sequence, and is prone to target drift or loss when the environment changes drastically or is occluded. This paper aims at the application background of the intelligent foot robot platform, where deep learning technology is used to adopt adaptive multi-target tracking. The frame detection template updated Shuffle net V2–0.5 convolutional neural network builds a deep tracking model, which speeds up the model calculation. At the same time, the multi-template input ensures that the required target information can be located in a larger search image and a global search module is added. The target position re-detection is carried out, and the background enhancement training is integrated to significantly strengthen the discrimination ability of the global search network. The target tracking accuracy of the improved visual object tracking algorithm reaches 64.7%, and the accuracy of the target located at the center point of the marker frame reaches 86.8%, which is significantly improved compared with the traditional algorithm.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bao, H., Lu, Y., Wang, Q.:Single target tracking via correlation filter and context adaptively. Multimedia Tools and Appl. 79, 27465–27482 (2020). https://doi.org/10.1007/s11042-020-09309-3
Wang, D., et al.: Online single target tracking in WAMI: benchmark and evaluation. Multimedia Tools Appl. 77(9), 10939–10960 (2018)
Xiao, J., et al.: Dynamic multi-level appearance models and adaptive clustered decision trees for single target tracking. Pattern Recognition 69.(2017). https://doi.org/10.1016/j.patcog.2017.04.001. Author, F.: Contribution title. In: 9th International Proceedings on Proceedings, pp. 1–2. Publisher, Location (2010)
Yanqing, W., Liang, Z., Cheng, X.: Fast target tracking based on improved deep sort and YOLOv3 fusion algorithm. Abstracts of the 7th International Conference of Pioneering Computer Scientists, Engineers and Educators (ICPCSEE 2021) Part I.Ed.. Springer, pp. 107–109 (2021). https://doi.org/10.1007/978-981-16-5940-9_27
Kwa, H.L., et al.: Optimal swarm strategy for dynamic target search and tracking. Autonomous Agents and MultiAgent Systems.Ed., pp. 672680 (2020)
Yıldırım, S., Jiang, L., Singh, S.S., Dean, T.A.: Calibrating the Gaussian multi-target tracking model. Stat. Comput. 25(3), 595–608 (2014)
Nam, H., Han, B.: Learning multi-domain convolutional neural networks for visual tracking. Computer Vision and Pattern Recognition IEEE (2015)
Tao, R., Gavves, E., Smeulders, A.: Siamese instance search for tracking. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1420–1429 (2016)
Bertinetto, L., et al.: Fully-Convolutional Siamese Networks for Object Tracking. CoRR abs/1606.09549 (2016)
Li, B., et al.: SiamRPN++: Evolution of siamese visual tracking with very deep networks. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) IEEE (2020)
Chen, Z.D., Zhong, B.N., Li, G.R., et al.: Siamese box adaptive network for visual tracking. In: Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle: IEEE, pp. 6667–6676 (2020)
Voigtlaender, P., Luiten, J., Torr, P.H.S., et al.: Siam R-CNN:Visual tracking by re-detection. In: Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, pp. 6577–6587 (2020)
Zhang, X., et al.: ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices. CoRR abs/1707.01083 (2017)
Grimaldi, M., et al.: Dynamic ConvNets on Tiny Devices via Nested Sparsity. arXiv e-prints (2022)
Sharma, S.: Ermenegildo Zegna OTB Process Analysis. (2015)
Bo, L., et al.: High performance visual tracking with siamese region proposal network. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) IEEE (2018)
Folberth, J., Becker, S.: Efficient Adjoint Computation for Wavelet and Convolution Operators (2017)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Zeng, L., He, W., Zhang, W. (2023). Visual Object Tracking with Adaptive Template Update and Global Search Augmentation. In: Yu, Z., Hei, X., Li, D., Song, X., Lu, Z. (eds) Intelligent Robotics. CIRAC 2022. Communications in Computer and Information Science, vol 1770. Springer, Singapore. https://doi.org/10.1007/978-981-99-0301-6_3
Download citation
DOI: https://doi.org/10.1007/978-981-99-0301-6_3
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-99-0300-9
Online ISBN: 978-981-99-0301-6
eBook Packages: Computer ScienceComputer Science (R0)