Skip to content
GitLab
Projects
Groups
Snippets
/
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
bob
bob.ip.qualitymeasure
Commits
5e724c93
Commit
5e724c93
authored
Mar 10, 2017
by
Sushil BHATTACHARJEE
Browse files
fixed print statements for python3
parent
f5f490b8
Pipeline
#7729
failed with stages
in 4 minutes and 2 seconds
Changes
2
Pipelines
1
Hide whitespace changes
Inline
Side-by-side
bob/ip/qualitymeasure/galbally_iqm_features.py
View file @
5e724c93
...
...
@@ -308,7 +308,7 @@ def vif(refImage, testImage):
#sc is scale, taking values (1,2,3,4)
for
sc
in
range
(
1
,
5
):
N
=
(
2
**
(
4
-
sc
+
1
))
+
1
#print
N, sc
#print
(
N, sc
)
win
=
gauss_2d
((
N
,
N
),
(
float
(
N
)
/
5.0
))
#ssg is scipy.signal
...
...
@@ -378,17 +378,17 @@ def high_low_freq_index(imgFFT, ncols):
fftRes
=
imgFFT
#np.fft.fft2(image)
fftMag
=
np
.
abs
(
fftRes
)
totalEnergy
=
np
.
sum
(
fftMag
)
#print
totalEnergy
#print
(
totalEnergy
)
lowIdx
=
colHalf
-
lowFreqColHalf
hiIdx
=
colHalf
+
lowFreqColHalf
#print
lowIdx, hiIdx
#print
(
lowIdx, hiIdx
)
LowFreqMag
=
fftMag
[:,
lowIdx
:
hiIdx
]
lowFreqMagTotal
=
np
.
sum
(
LowFreqMag
)
fftMag
[:,
lowIdx
:
hiIdx
]
=
0
highFreqMagTotal
=
np
.
sum
(
fftMag
)
#print
'partial freq. sums:
',
lowFreqMagTotal, highFreqMagTotal
#print
(
'partial freq. sums:
%f %f' %(
lowFreqMagTotal, highFreqMagTotal
))
highLowFreqIQ
=
np
.
abs
(
lowFreqMagTotal
-
highFreqMagTotal
)
/
float
(
totalEnergy
)
...
...
@@ -447,8 +447,8 @@ def testRegionalMax():
A
[
2
,
7
]
=
45
A
[
3
,
8
]
=
44
rm
=
regionalmax
(
A
)
print
A
print
rm
print
(
A
)
print
(
rm
)
"""
find local maxima using 3x3 mask.
...
...
@@ -536,25 +536,6 @@ def cornerMetric(image):
return
cornerness
#
# def imshow(image):
# import matplotlib
# from matplotlib import pyplot as plt
# if len(image.shape)==3:
# #imshow() expects color image in a slightly different format, so first rearrange the 3d data for imshow...
# outImg = image.tolist()
# print len(outImg)
# result = np.dstack((outImg[0], outImg[1]))
# outImg = np.dstack((result, outImg[2]))
# plt.imshow((outImg*255.0).astype(np.uint8)) #[:,:,1], cmap=mpl.cm.gray)
#
# else:
# if(len(image.shape)==2):
# #display gray image.
# plt.imshow(image.astype(np.uint8), cmap=matplotlib.cm.gray)
#
# plt.show()
'''
compute the corner-based similarity between 2 images (how close are the numbers of corners found in the two images?).
returns an index between 0 and 1. The smaller the better.
...
...
@@ -574,7 +555,7 @@ def corner_similarity(refImage, testImage):
nCornersRef
=
np
.
sum
(
C_peaks
)
nCornersTest
=
np
.
sum
(
CG_peaks
)
#print
'CornerSim::', nCornersRef, nCornersTest
#print
(
'CornerSim::
%f %f
',
%(
nCornersRef, nCornersTest
) )
maxCorners
=
max
(
nCornersRef
,
nCornersTest
)
...
...
@@ -633,8 +614,6 @@ def edge_thinning(bx, by, thinning=1):
n
=
by
.
shape
[
1
]
e
=
np
.
zeros
([
m
,
n
],
dtype
=
np
.
uint8
)
# will contain the resulting edge-map.
# print 'bx', bx.shape
# print 'by', by.shape
#compute the edge-strength from the 2 directional filter-responses
b
=
np
.
sqrt
(
bx
*
bx
+
by
*
by
)
...
...
@@ -807,7 +786,7 @@ def gaussianSmooth(image):
# def testQualityMeasures(image, smoothed):
# frameFeatSet = imageQualityMeasures(image, smoothed)
#
# print
frameFeatSet
# print
(
frameFeatSet
)
#
#
# if __name__ == '__main__':
...
...
@@ -817,8 +796,8 @@ def gaussianSmooth(image):
# diffImg = image - smoothed
# #test MAMS
# mas, mams = angle_similarity2(image, smoothed, diffImg)
# print
'mas
',
mas
# print
'mams
',
mams
# print
(
'mas
: %f' %
mas
)
# print
(
'mams
: %f' %
mams
)
...
...
bob/ip/qualitymeasure/msu_iqa_features.py
View file @
5e724c93
...
...
@@ -59,7 +59,7 @@ def imshow(image):
if
len
(
image
.
shape
)
==
3
:
#imshow() expects color image in a slightly different format, so first rearrange the 3d data for imshow...
outImg
=
image
.
tolist
()
print
len
(
outImg
)
print
(
len
(
outImg
)
)
result
=
np
.
dstack
((
outImg
[
0
],
outImg
[
1
]))
outImg
=
np
.
dstack
((
result
,
outImg
[
2
]))
plt
.
imshow
((
outImg
*
255.0
).
astype
(
np
.
uint8
))
#[:,:,1], cmap=mpl.cm.gray)
...
...
@@ -215,19 +215,19 @@ def marzilianoBlur(image):
# ind = np.lexsort((row,col))
# row = row[ind]
# col = col[ind]
#print
'lexsort_col:
',
1+col
#print
'lexsort_row:
',
1+row
#print
(
'lexsort_col:
%d' % (
1+col
))
#print
(
'lexsort_row:
%d' % (
1+row
))
#This was only used for debugging (to compare with Matlab code). In fact it is not necessary, so it is commented out.
edgeWidths
=
np
.
zeros_like
(
row
,
dtype
=
int
)
firstRow
=
row
[
0
]
# print
'firstRow:
',
firstRow
# print
(
'firstRow:
%d' %
firstRow
)
for
i
in
range
(
len
(
row
)):
rEdge
=
row
[
i
]
cEdge
=
col
[
i
]
# if rEdge == firstRow: print
"edgePoints:
",
(i, rEdge, cEdge)
# if rEdge == firstRow: print
(
"edgePoints:
%d %d %d" %
(i, rEdge, cEdge)
)
cStart
=
0
# instead of setting them to 'inf' as in MSU's Matlab version
cEnd
=
0
...
...
@@ -288,22 +288,20 @@ def calmoment( channel, regionMask=None ):
m
=
np
.
mean
(
channel
)
# m = mean(channel(:));
d
=
np
.
std
(
channel
)
# d = sqrt(sum((channel(:) - m) .^ 2) / pixnum);
s
=
np
.
sum
(
np
.
power
(
((
channel
-
m
)
/
d
),
3
))
/
nPix
# s = sum(((channel(:) - m) ./ d) .^ 3) / pixnum;
#print
't:', t
#print
(t)
myHH
=
np
.
histogram
(
channel
,
t
)[
0
]
#print
myHH
#print
(
myHH
)
hh
=
myHH
.
astype
(
float
)
/
nPix
# hh = hist(channel(:),t) / pixnum;
#print 'numPix:', nPix
#print 'histogram:',hh
#print('numPix: %d' % nPix)
#H = np.array([m,d,s, np.sum(hh[0:1]), np.sum(hh[-2:-1])]) # H = [m d s sum(hh(1:2)) sum(hh(end-1:end))];
H
=
np
.
array
([
m
,
d
,
s
])
s0
=
np
.
sum
(
hh
[
0
:
2
])
#print
s0
#print
(
s0
)
H
=
np
.
hstack
((
H
,
s0
))
s1
=
np
.
sum
(
hh
[
-
2
:])
#print
s1
#print
(
s1
)
H
=
np
.
hstack
((
H
,
s1
)
)
#print 'calmoment:',H.shape
return
H
...
...
@@ -343,7 +341,6 @@ def calColorHist(image, m=100):
#1. compute color histogram of image (normalized, if specified)
numBins
=
32
maxval
=
255
#print "calColorHist():: ", image.shape
cHist
=
rgbhist
(
image
,
maxval
,
numBins
,
1
)
#2. determine top 100 colors of image from histogram
...
...
@@ -352,10 +349,8 @@ def calColorHist(image, m=100):
cHist
=
y
[
0
:
m
]
# H = Y(1:m)';
c
=
np
.
cumsum
(
y
)
# C = cumsum(Y);
# print 'cumsum shape:', c.shape
# thresholdedC = np.where(c>0.999)
# # print thresholdedC.shape
# print 'thresholdedC:', thresholdedC[0][0] #:200]
# print('thresholdedC: %f' % thresholdedC[0][0]) #:200]
numClrs
=
np
.
where
(
c
>
0.99
)[
0
][
0
]
# clrnum = find(C>.99,1,'first') - 1;
cHist
=
np
.
array
(
cHist
)
...
...
@@ -371,17 +366,16 @@ def rgbhist(image, maxval, nBins, normType=0):
H
=
np
.
zeros
((
nBins
,
nBins
,
nBins
),
dtype
=
np
.
uint32
)
# zeros([nBins nBins nBins]);
# testImage = image[:,0:3,0:3].copy()
# print
testImage.shape
# print
image.shape
# print
testImage[0,:,:]
# print
''
# print
testImage[1,:,:]
# print
''
# print
testImage[2,:,:]
# print
''
# print
testImage.reshape(3, 9, order='C').T
# print
(
testImage.shape
)
# print
(
image.shape
)
# print
(
testImage[0,:,:]
)
# print
(
''
)
# print
(
testImage[1,:,:]
)
# print
(
''
)
# print
(
testImage[2,:,:]
)
# print
(
''
)
# print
(
testImage.reshape(3, 9, order='C').T
)
#
# assert(0>1), "Stop!"
decimator
=
(
maxval
+
1
)
/
nBins
numPix
=
image
.
shape
[
1
]
*
image
.
shape
[
2
]
...
...
@@ -396,8 +390,8 @@ def rgbhist(image, maxval, nBins, normType=0):
#totalNBins = np.prod(H.shape)
#H = H.reshape(totalNBins, 1, order='F') same as H = reshape(H, nBins**3, 1)
H
=
H
.
ravel
()
#H = H(:);
# print
'H type:
',
type(H[0])
# print
H.shape
# print
(
'H type:
%s' %(
type(H[0])
)
# print
(
H.shape
)
# Un-Normalized histogram
if
normType
==
1
:
H
=
H
.
astype
(
np
.
float32
)
/
np
.
sum
(
H
)
# l1 normalization
...
...
Write
Preview
Supports
Markdown
0%
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment