• Acta Optica Sinica
  • Vol. 38, Issue 2, 0215003 (2018)
Haosheng Gao, Chaoying Tang*, Xiaoteng Chen, and Xiao Yu
Author Affiliations
  • College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 211106, China
  • show less
    DOI: 10.3788/AOS201838.0215003 Cite this Article Set citation alerts
    Haosheng Gao, Chaoying Tang, Xiaoteng Chen, Xiao Yu. Line Tracking Method of Arm Vein Based on Bayesian Theory[J]. Acta Optica Sinica, 2018, 38(2): 0215003 Copy Citation Text show less
    Determination of initial seed points. (a) NIR image; (b) image after Gabor filtering; (c) grid lines and candidate initial seed points; (d) effective initial seed points
    Fig. 1. Determination of initial seed points. (a) NIR image; (b) image after Gabor filtering; (c) grid lines and candidate initial seed points; (d) effective initial seed points
    Dynamic search area of kth iteration
    Fig. 2. Dynamic search area of kth iteration
    Partial initial seed points and their corresponding local vessel directions
    Fig. 3. Partial initial seed points and their corresponding local vessel directions
    Three types of vessel structures. (a) Normal; (b) bifurcation; (c) crossing
    Fig. 4. Three types of vessel structures. (a) Normal; (b) bifurcation; (c) crossing
    Gaussian model of vessel cross section
    Fig. 5. Gaussian model of vessel cross section
    dm1 and dm2 corresponding to normal vessel
    Fig. 6. dm1 and dm2 corresponding to normal vessel
    Tracking results. (a) NIR image; (b) image processed with Gabor filter; (c) tracked vessel edge points; (d) edge points of normal vessel; (e) edge points of bifurcation vessel; (f) edge points of crossing vessel
    Fig. 7. Tracking results. (a) NIR image; (b) image processed with Gabor filter; (c) tracked vessel edge points; (d) edge points of normal vessel; (e) edge points of bifurcation vessel; (f) edge points of crossing vessel
    Comparison of results of three methods. (a) NIR image 1; (b) NIR image 2; (c) image 1 processed with Gabor filter; (d) image 2 processed with Gabor filter; (e) image 1 processed with proposed method; (f) image 2 processed with proposed method; (g) image 1 processed with RLT; (h) image 2 processed with RLT; (i) image 1 processed with LAT; (j) image 2 processed with LAT
    Fig. 8. Comparison of results of three methods. (a) NIR image 1; (b) NIR image 2; (c) image 1 processed with Gabor filter; (d) image 2 processed with Gabor filter; (e) image 1 processed with proposed method; (f) image 2 processed with proposed method; (g) image 1 processed with RLT; (h) image 2 processed with RLT; (i) image 1 processed with LAT; (j) image 2 processed with LAT
    ROC images extracted by three methods. (a) Proposed method; (b) RLT method; (c) LAT method
    Fig. 9. ROC images extracted by three methods. (a) Proposed method; (b) RLT method; (c) LAT method
    Comparison of detection rate of three methods with different SNR. (a) Precision; (b) recall; (c) F-measure
    Fig. 10. Comparison of detection rate of three methods with different SNR. (a) Precision; (b) recall; (c) F-measure
    MethodRprecisionRrecallRF-measure
    Proposed method0.330.360.34
    RLT method0.350.300.32
    LAT method0.090.170.12
    Table 1. Comparison of detection rate of three methods
    Haosheng Gao, Chaoying Tang, Xiaoteng Chen, Xiao Yu. Line Tracking Method of Arm Vein Based on Bayesian Theory[J]. Acta Optica Sinica, 2018, 38(2): 0215003
    Download Citation