• Chinese Journal of Lasers
  • Vol. 49, Issue 20, 2007302 (2022)
Cong Chen1, Miao Liu1、*, Jigang Wang2, and Shourui Yang3
Author Affiliations
  • 1Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin 300072, China
  • 2Tianjin Haihe Hospital, Tianjin 300222, China
  • 3School of Computer Science and Engineering, Tianjin University of Technology, Tianjin 300384, China
  • show less
    DOI: 10.3788/CJL202249.2007302 Cite this Article Set citation alerts
    Cong Chen, Miao Liu, Jigang Wang, Shourui Yang. Spatial Registration Method for Neuronavigation Using Adaptive Thresholds[J]. Chinese Journal of Lasers, 2022, 49(20): 2007302 Copy Citation Text show less
    Flow chart of adaptive threshold registration method
    Fig. 1. Flow chart of adaptive threshold registration method
    Schematic of binocular stereo vision
    Fig. 2. Schematic of binocular stereo vision
    Picture of face point cloud collection system
    Fig. 3. Picture of face point cloud collection system
    Calculation diagram of fast point feature histogram (FPFH)
    Fig. 4. Calculation diagram of fast point feature histogram (FPFH)
    Head model
    Fig. 5. Head model
    Target balls and their labels
    Fig. 6. Target balls and their labels
    Three-dimensional reconstruction result of CT images. (a) Point cloud of face;(b) point cloud of target balls
    Fig. 7. Three-dimensional reconstruction result of CT images. (a) Point cloud of face;(b) point cloud of target balls
    Point clouds of face in CT and world coordinate system. (a) CT point cloud; (b) point cloud 0; (c) point cloud 1; (d) point cloud 2; (e) point cloud 3
    Fig. 8. Point clouds of face in CT and world coordinate system. (a) CT point cloud; (b) point cloud 0; (c) point cloud 1; (d) point cloud 2; (e) point cloud 3
    Point clouds of face after downsampling. (a) CT point cloud; (b) point cloud 0; (c) point cloud 1; (d) point cloud 2; (e) point cloud 3
    Fig. 9. Point clouds of face after downsampling. (a) CT point cloud; (b) point cloud 0; (c) point cloud 1; (d) point cloud 2; (e) point cloud 3
    Residual distributions of coarse registration. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Fig. 10. Residual distributions of coarse registration. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Counts of each points excluded. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Fig. 11. Counts of each points excluded. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Residual distribution of fine registration of general ICP method. (a) No deformation; (b) deformation 1;
    Fig. 12. Residual distribution of fine registration of general ICP method. (a) No deformation; (b) deformation 1;
    Residual distribution of fine registration of our method. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Fig. 13. Residual distribution of fine registration of our method. (a) No deformation; (b) deformation 1; (c) deformation 2; (d) deformation 3
    Box plot of mean registration error of two methods
    Fig. 14. Box plot of mean registration error of two methods
    ConditionTx/mmTy/mmTz/mm
    No deformation-161.94 to -148.02-621.58 to -614.03-13.27 to 6.82
    Deformation 1-162.86 to -146.08-621.93 to -612.69-25.51 to 15.32
    Deformation 2-174.79 to -134.04-627.82 to -607.69-16.04 to 9.02
    Deformation 3-158.32 to -136.13-644.67 to -635.54-93.69 to -65.00
    Table 1. Transformation parameter range of coarse registration
    Target labelError /mm
    No deformationDeformation 1Deformation 2Deformation 3
    10.75±0.050.34±0.120.33±0.100.67±0.16
    20.60±0.040.37±0.110.25±0.040.56±0.20
    30.67±0.030.32±0.080.29±0.060.61±0.13
    40.58±0.030.44±0.060.23±0.030.63±0.17
    50.54±0.030.34±0.070.28±0.040.51±0.18
    60.45±0.030.35±0.050.23±0.050.50±0.17
    70.68±0.030.32±0.060.19±0.030.71±0.10
    80.53±0.040.25±0.080.27±0.040.64±0.13
    110.57±0.050.37±0.120.29±0.060.65±0.15
    120.43±0.060.32±0.130.35±0.080.60±0.14
    130.40±0.050.27±0.070.24±0.090.57±0.11
    140.52±0.070.29±0.140.26±0.060.26±0.20
    150.48±0.070.42±0.170.41±0.070.38±0.14
    Mean value0.55±0.050.34±0.100.28±0.060.56±0.15
    Table 2. Target registration error of our method
    Target labelError /mm
    No deformationDeformation 1Deformation 2Deformation 3
    10.74±0.051.81±0.052.64±0.050.96±0.11
    20.60±0.041.66±0.052.41±0.050.56±0.14
    30.67±0.031.97±0.043.02±0.030.95±0.24
    40.57±0.031.83±0.042.65±0.030.66±0.07
    50.53±0.031.65±0.042.43±0.030.65±0.08
    60.45±0.031.58±0.042.22±0.030.64±0.11
    70.68±0.032.06±0.032.98±0.031.13±0.21
    80.53±0.031.89±0.032.68±0.030.84±0.15
    110.57±0.041.87±0.032.67±0.041.09±0.07
    120.41±0.051.72±0.042.53±0.041.16±0.10
    130.39±0.041.77±0.052.17±0.051.15±0.09
    140.50±0.061.96±0.082.76±0.061.71±0.08
    150.47±0.061.77±0.072.55±0.061.64±0.09
    Mean value0.55±0.041.81±0.052.59±0.041.01±0.12
    Table 3. Target registration error of general ICP method
    Cong Chen, Miao Liu, Jigang Wang, Shourui Yang. Spatial Registration Method for Neuronavigation Using Adaptive Thresholds[J]. Chinese Journal of Lasers, 2022, 49(20): 2007302
    Download Citation