Enhanced Region Growing for Brain Tumor MR Image Segmentation

. 2021 Feb 1;7(2):22. doi: 10.3390/jimaging7020022

Algorithm 2 Enhanced Region Growing Segmentation for Brain Tumor Segmentation

1:
input: skull stripped image, $i m$
2:
Resize the Image

$i m \leftarrow i m r e s i z e (i m, [256, 256])$
3:
iterate through each $8 \times 8$ block
4:
for $i = 1 : 8 : 256$ do
5:
for $j = 1 : 8 : 256$ do
6:
Collect the mean of each block

$m I s \leftarrow m e a n (i m (i : i + 7, j : j + 7)$
7:
Collect the centers of each block

$c B s \leftarrow [i + 3, j + 3]$
8:
end for
9:
end for
10:
Select top 5 blocks based on the intensity

$[i n d, v a l s] = m a x (m I s, 5)$

$s e e d s = c B s (i n d)$
11:
return $s e e d s$
12:
for $m = 1 : 5$ do
13:
ROI $_{m}$ = Region-growing(seed $_{m}$ )
14:
end for
15:
Compare each ROI $_{m}$ against GT using evaluation parameters for $m = 1 : 5$
16:
Select the best ROI as a final segmentation output.