Online learning discriminative sparse convolution networks for robust UAV object tracking

被引:1
作者
Xu, Qi [1 ]
Xu, Zhuoming [1 ]
Wang, Huabin [2 ]
Chen, Yun [1 ]
Tao, Liang [2 ]
机构
[1] Hohai Univ, Coll Comp Sci & Software Engn, Nanjing 210098, Peoples R China
[2] Anhui Univ, Sch Comp Sci & Technol, Hefei 230031, Peoples R China
关键词
UAV object tracking; Online learning; Deep learning; Convolutional networks; Sparse constrains;
D O I
10.1016/j.knosys.2024.112742
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Despite the remarkable empirical success for UAV object tracking, current convolutional networks usually have three unavoidable limitations: (1) The feature maps produced by convolutional layers are difficult to interpret. (2) The network needs to be trained offline on a large-scale auxiliary training set, resulting in the feature extraction ability of the trained network depending on the categories of the training set. (3) The performance of networks suffers from sensitivity to hyper-parameters (such as learning rate and weight decay) when the network needs online fine-tuning. To overcome the three limitations, this paper proposes a Discriminative Sparse Convolutional Network (DSCN) that exhibits good layer-wise interpretability and can be trained online without requiring any auxiliary training data. By imposing sparsity constraints on the convolutional kernels, DSCN furnishes the convolution layer with an explicit data meaning, thus enhancing the interpretability of the feature maps. These convolutional kernels are directly learned online from image blocks, which eliminates the offline training process on auxiliary data sets. Moreover, a simple yet effective online tuning method with few hyper-parameters is proposed to fine-tune fully connected layers online. We have successfully applied DSCN to UAV object tracking and conducted extensive experiments on six mainstream UAV datasets. The experimental results demonstrate that our method performs favorably against several state-of-the-art tracking algorithms in terms of tracking accuracy and robustness.
引用
收藏
页数:18
相关论文
共 50 条
  • [41] Long-term object tracking combined offline with online learning
    Hu, Mengjie
    Wei, Zhenzhong
    Zhang, Guangjun
    [J]. OPTICAL ENGINEERING, 2016, 55 (04)
  • [42] CIOFL: Collaborative Inference- Based Online Federated Learning for UAV Object Detection
    Wu, Feiyu
    Dong, Chao
    Qu, Yuben
    Sun, Hao
    Zhang, Lei
    Wu, Qihui
    [J]. 2022 IEEE 19TH INTERNATIONAL CONFERENCE ON MOBILE AD HOC AND SMART SYSTEMS (MASS 2022), 2022, : 258 - 259
  • [43] Object tracking based on an online learning network with total error rate minimization
    Jang, Se-In
    Choi, Kwontaeg
    Toh, Kar-Ann
    Teoh, Andrew Beng Jin
    Kim, Jaihie
    [J]. PATTERN RECOGNITION, 2015, 48 (01) : 126 - 139
  • [44] Speeded Up Low-Rank Online Metric Learning for Object Tracking
    Cong, Yang
    Fan, Baojie
    Liu, Ji
    Luo, Jiebo
    Yu, Haibin
    [J]. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2015, 25 (06) : 922 - 934
  • [45] LEARNING HIERARCHICAL FEATURES FOR VISUAL OBJECT TRACKING WITH RECURSIVE NEURAL NETWORKS
    Wang, Li
    Liu, Ting
    Wang, Bing
    Lin, Jie
    Yang, Xulei
    Wang, Gang
    [J]. 2019 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2019, : 3088 - 3092
  • [46] Deep Learning-Based Multi-class Multiple Object Tracking in UAV Video
    Micheal, A. Ancy
    Vani, K.
    [J]. JOURNAL OF THE INDIAN SOCIETY OF REMOTE SENSING, 2022, 50 (12) : 2543 - 2552
  • [47] Deep Learning-Based Multi-class Multiple Object Tracking in UAV Video
    A. Ancy Micheal
    K. Vani
    [J]. Journal of the Indian Society of Remote Sensing, 2022, 50 : 2543 - 2552
  • [48] Research progress of UAV aerial video multi⁃object detection and tracking based on deep learning
    Yuan Y.
    Wu Y.
    Zhao L.
    Chen J.
    Zhao Q.
    [J]. Hangkong Xuebao/Acta Aeronautica et Astronautica Sinica, 2023, 44 (18):
  • [49] Cutting temperature field online reconstruction using temporal convolution and deep learning networks
    Zheng, Yitong
    Yin, Zengbin
    [J]. INTERNATIONAL JOURNAL OF HEAT AND MASS TRANSFER, 2025, 241
  • [50] Kernel-based Online Object Tracking via Gaussian Mixture Model Learning
    Miao, Quan
    Gu, Yanfeng
    [J]. PROCEEDINGS OF 2016 SIXTH INTERNATIONAL CONFERENCE ON INSTRUMENTATION & MEASUREMENT, COMPUTER, COMMUNICATION AND CONTROL (IMCCC 2016), 2016, : 522 - 525