Pattern Classification Methods for Analysis and Visualization of Brain Perfusion CT Maps - Computational Intelligence Paradigms in Advanced Pattern Classification - page 156

Table 3 Image registration algorithms and their parameters (

σ

is a value of variation

coefficient in Gaussian kernel)

Algorithm
designation

Registration algorithm

A Affine algorithm

B

σ = 0 . 5

TDA, , 2000 iterations

C

σ = 1

TDA, , 2000 iterations

D

σ = 2

TDA, , 2000 iterations

E

σ = 5

TDA, , 2000 iterations

F

σ = 10

TDA, , 2000 iterations

σ = 15

G

TDA, , 2000 iterations

H FFD, initial grid size 6x6, 12 iterations

I FFD, initial grid size 8x8, 12 iterations

J FFD, initial grid size 10x10, 12 iterations

K FFD, initial grid size 16x16, 12 iterations

L FFD, initial grid size 32x32, 12 iterations

M FFD, initial grid size 6x6, 12 iterations, add

N FFD, initial grid size 8x8, 12 iterations, affine rotation and scaling

O FFD, initial grid size 10x10, 12 iterations, affine rotation and scaling

P FFD, initial grid size 16x16, 12 iterations, affine rotation and scaling

Q FFD, initial grid size 32x32, 12 iterations, affine rotation and scaling

R FFD, initial grid size 6x6, 12 iterations, 1 mesh refinement

S FFD, initial grid size 8x8, 12 iterations, 1 mesh refinement

T FFD, initial grid size 10x10, 12 iterations, 1 mesh refinement

U FFD, initial grid size 6x6, 12 iterations, 1 mesh refinement, affine rotation

and scaling

V FFD, initial grid size 8x8, 12 iterations, 1 mesh refinement, affine rotation

and scaling

W FFD, initial grid size 10x10, 12 iterations, 1 mesh refinement, affine rota-

tion and scaling

X FFD, initial grid size 6x6, 12 iterations, 2 mesh refinement

Y FFD, initial grid size 8x8, 12 iterations, 2 mesh refinement

Z FFD, initial grid size 10x10, 12 iterations, 2 mesh refinement

1 FFD, initial grid size 6x6, 12 iterations, 2 mesh refinement, affine rotation

and scaling

2 FFD, initial grid size 8x8, 12 iterations, 2 mesh refinement, affine rotation

and scaling

3 FFD, initial grid size 10x10, 12 iterations, 2 mesh refinement, affine rota-

tion and scaling