A Locally Linear Method for Enforcing Temporal Smoothness in Serial Image Registration - Longitudinal Registration and Shape Modeling - page 17

Table 1. Results of the proposed and reference methods when registering two full
synthetic time-series ( Dense Sampling ) and reconstructing unseen time-points from
sparsely sampled time-seris ( Interpolation )

Dense Sampling Interpolation
pair-wise uniform adaptive pair-wise uniform adaptive

Single Sulcus
Mean Error

23.83 19.25 22.95 86.64 134.14 84.94

Smoothness 40.06 41.27 28.56 68.35 71.56 55.34

Two Sulci
Mean Error 40.45 23.13 37.56 128.92 130.90 108.30

Smoothness 57.05 42.64 39.69 119.73 111.94 111.33

70

120

60

100

50

80

40

60

30

40

20

20

10

0 0

Time Time

(a) Single Sulcification (b) Pairwise Sulcification

Fig. 5. Registration error using full image data (orange: no, blue: uniform, green:
spatially adaptive temporal smoothing)

200 700

600

150

500

400

100

300

200

50

100

0 0

Time Time

(a) Single Sulcification (b) Pairwise Sulcification

Fig. 6. Smoothness of deformation ∇ det( v t ( x )) ,x ∈ C (orange: no, blue: uniform,

green: spatially adaptive temporal smoothing)