376
F. Shi et al.
(a)
(b)
(c)
y = 0.9245x + 0.0975
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
GT2
GT1
3
mm
3
mm
y = 1.0186x + 0.0079
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
Results
GT1
3
mm
3
mm
y = 0.9676x + 0.0686
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
GT2
Results
3
mm
3
mm
Fig. 13.8 a, b and c represent the correlation of ground truth I with ground truth II, the correlation
of ground truth I with the predicted results and the correlation of ground truth II with the predicted
results, respectively
(a)
(b)
Fig. 13.9 a and b represent CNV growth parameters curve of treatment group and reference group
each curve represents the estimated parameter for the last time point. Figure 13.10
shows two examples of results of CNV growth prediction, compared with the ground
truth. The TPVF, FPVF and Dice coefficients (DC) for each subject are listed in
Table 13.1. The mean values of TPVF, FPVF and DC are 78.41, 2.44 and 79.22%,
respectively, which shows the efficacy of the proposed method to predict the future
status of CNV.
F. Shi et al.
(a)
(b)
(c)
y = 0.9245x + 0.0975
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
GT2
GT1
3
mm
3
mm
y = 1.0186x + 0.0079
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
Results
GT1
3
mm
3
mm
y = 0.9676x + 0.0686
0
0.5
1
1.5
2
2.5
3
3.5
4
0
0.5
1
1.5
2
2.5
3
3.5 4
GT2
Results
3
mm
3
mm
Fig. 13.8 a, b and c represent the correlation of ground truth I with ground truth II, the correlation
of ground truth I with the predicted results and the correlation of ground truth II with the predicted
results, respectively
(a)
(b)
Fig. 13.9 a and b represent CNV growth parameters curve of treatment group and reference group
each curve represents the estimated parameter for the last time point. Figure 13.10
shows two examples of results of CNV growth prediction, compared with the ground
truth. The TPVF, FPVF and Dice coefficients (DC) for each subject are listed in
Table 13.1. The mean values of TPVF, FPVF and DC are 78.41, 2.44 and 79.22%,
respectively, which shows the efficacy of the proposed method to predict the future
status of CNV.
