Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 57, Issue 18, 181025(2020)

Residual Network Feature Fusion Tracking Algorithm Based on Graph Salience Detection

Hui Jin1,2 and Xinyang Li1、*
Author Affiliations
  • 1Key Laboratory on Adaptive Optics, Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu, Sichuan 610209, China
  • 2University of Chinese Academy of Sciences, Beijing 100049, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (9)
    Equations (0)
    References (19)
    Cited By (1)
    Tools
    Paper Information
    Topics
    Figures & Tables(9)
    Steps of target location

    Fig. 1. Steps of target location

    Download full size
    Original image and GBVS image. (a) Original image; (b) GBVS image

    Fig. 2. Original image and GBVS image. (a) Original image; (b) GBVS image

    Download full size
    Improved flow chart based on GBVS

    Fig. 3. Improved flow chart based on GBVS

    Download full size
    Average center error of the improved algorithm based on GBVS

    Fig. 4. Average center error of the improved algorithm based on GBVS

    Download full size
    Tracking effect of the algorithm. (a) Before improvement; (b) after improvement

    Fig. 5. Tracking effect of the algorithm. (a) Before improvement; (b) after improvement

    Download full size
    Tracking accuracy of different algorithms on the OTB50 data set. (a) Total accuracy; (b) fast movement; (c) motion blur; (d) low resolution; (e) occlusion; (f) different scales

    Fig. 6. Tracking accuracy of different algorithms on the OTB50 data set. (a) Total accuracy; (b) fast movement; (c) motion blur; (d) low resolution; (e) occlusion; (f) different scales

    Download full size
    Tracking results of different algorithms in the OTB dataset

    Fig. 7. Tracking results of different algorithms in the OTB dataset

    Download full size

    • Table 1. Comparison of average center error based on GBVS improved algorithm

      View table

      Table 1. Comparison of average center error based on GBVS improved algorithm

      Layer ofResNet-50 (number)Conv1-1(2)Addition2-1(6)Conv2-3(12)Addition2-2(26)Conv2-10(34)Addition3-1(48)
      After GBVS16.918.2812.189.304.917.43
      Before GBVS18.1510.2214.3910.786.689.55
      Error1.241.942.211.481.772.12

    • Table 2. Average central error of different algorithms

      View table

      Table 2. Average central error of different algorithms

      AlgorithmSTRCFStapleCSKBeforeGBVSAfterGBVS
      Center error6.7955.1428.1886.6864.916
    Tools

    Get Citation

    Copy Citation Text

    Hui Jin, Xinyang Li. Residual Network Feature Fusion Tracking Algorithm Based on Graph Salience Detection[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181025

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Image Processing

    Received: Feb. 5, 2020

    Accepted: Mar. 25, 2020

    Published Online: Sep. 2, 2020

    The Author Email: Xinyang Li (xyli@ioe.ac.cn)

    DOI:10.3788/LOP57.181025

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology