1(d)
for i=1:6
outclip= adapthisteq(brain,'clipLimit',i*.01);
outtile= adapthisteq(brain,'NumTiles',[4*(2^i) 4*(2^i)]);
end
1000
700
700
600
800
600
500
500
600
400
400
400
300
300
200
200
100
100
200
0
0
0
50
100
150
200
250
0
0
50
100
150
200
250
700
700
600
600
600
500
500
500
400
400
400
300
300
300
200
200
200
100
100
100
0
0
0
50
100
150
200
250
0
50
100
150
200
250
0
50
100
150
200
250
0
0
50
100
150
200
250
AS increasing “ClipLimit”, the cortex in the foreground gets clearer and shows more
detail. If we keep increasing “ClipLimit”, the noise in the dark background is
enhanced too much that the whole image starts to look messy. The parameter
“ClipLimit” set a limit for the contrast enhancement. The higher value of “ClipLimit”
gives higher contrast and a more flat histogram. The Maximum value of ClipLimit=1
results in standard adaptive histogram equalization with no clipping. To brighten the
foreground and prevent a too noisy background, the best value of ClipLimit should be
around 0.02~0.03.
Then we increase “NumTiles” by powers of 2. Higher value of “NumTiles” divides
the image into a smaller region for local histogram equalization. As we increase
“NumTiles”, it gives a better contrast in the foreground, and the best value can be [16
16] or [32 32]. However, when the value is further increased, the number of the data
point in each tile would be too few for the equalization to do a good adjustment.