Surface registration by markers guided nonrigid iterative closest points algorithm - Emerging Trends in Image Processing, Computer Vision, and Pattern Recognition - page 413

ure 1 . As markers 15 mm white squares atached to the abdomen were used, which corners
were manually segmented by two users.

FIGURE 1 Example of input data: intensity map for ToF camera of abdomen with nine square
markers.

3 Results
For the different methods of finding correspondences, evaluation scores: surface distances,
correspondence distances and average marker error, are presented in Tables 1 and 2 .

Table 1
Surface distances for four variants of ICP

Surface Distance [mm]

ID

Initial E NH SM DM

1 5.83 0,21 0.6 0.69 0.63

2 12.04 0.12 0.74 0.87 0.61

3 25.16 0.03 0.03 0.06 3.43

4 19.84 0.14 1.04 0.51 3.56

5 5.21 0.004 0.21 0.28 0.28

6 10.76 0.16 0.15 0.09 1.94

E, the Euclidean distance; NH, normal vectors with initial rigid registration; SM, static; DM, dynamic markers vectors.