-
Notifications
You must be signed in to change notification settings - Fork 1
/
createfeatures.py
199 lines (173 loc) · 7.56 KB
/
createfeatures.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
from saliencymap import Dij, estimateCenter, gD, gauss
from localmax import getLocalMax, convertIndices
import matplotlib.pyplot as plt
import cv2
from os.path import isfile, join
from os import listdir
from computecorners import makePairs, affineTransform
from parameters import setHoG, getParameters, showHoG
import numpy as np
from matplotlib import pyplot as plt
from numpy import arange, sqrt, exp, pi, meshgrid, arctan, zeros, ceil
from scipy.ndimage import convolve1d, convolve
from scipy.misc import imread
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from os.path import isfile, join
from os import listdir
from skimage import io, color
from skimage.feature import hog
from skimage import data, color, exposure
def getImages(mypath):
onlyfiles = [f for f in listdir(mypath) if isfile(join(mypath,f))]
images = np.empty(len(onlyfiles), dtype=object)
for n in range(0, len(onlyfiles)):
images[n]= cv2.imread(join(mypath, onlyfiles[n]))
return images
def getSaliency(img, firstKernel, secondKernel):
Dx,Dy = Dij(img, firstKernel) # get displacement x and y
# get image with saliency
saliency, blurredSaliency = estimateCenter(Dx,Dy, secondKernel)
return blurredSaliency
def displayThis(image):
fd, hog_image = hog(image, orientations=8, pixels_per_cell=(16, 16),
cells_per_block=(1, 1), visualise=True)
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(8, 4), sharex=True, sharey=True)
ax1.axis('off')
ax1.imshow(image, cmap=plt.cm.gray)
ax1.set_title('Input image')
ax1.set_adjustable('box-forced')
# Rescale histogram for better display
hog_image_rescaled = exposure.rescale_intensity(hog_image, in_range=(0, 0.02))
ax2.axis('off')
ax2.imshow(hog_image_rescaled, cmap=plt.cm.gray)
ax2.set_title('Histogram of Oriented Gradients')
ax1.set_adjustable('box-forced')
plt.show()
def extractFeatures(image, indices, kernel, imageNum, M, N):
bs = 200
borderReplicate = cv2.copyMakeBorder(image, bs,bs,bs,bs,cv2.BORDER_REPLICATE)
newIndices = convertIndices(indices,bs)
pairs = makePairs(newIndices) # make pairs of all the indices
parametersList = np.array([])
hight = img.shape[0]
width = img.shape[1]
#print img.shape
#print len(pairs)
amount = len(pairs)
for p in range(0, amount):
pair = pairs[p]
#for pair in pairs:
# possible = True # default is true
# check if affine transformation is possible
#for coordinates in pair:
# if coordinates[0] < 0 or coordinates[0] > hight or coordinates[1] < 0 or coordinates[1] > width:
# possible = False # if coordinate is negative set to false
# # break if one coordinate is negative, no need to check the rest
# print pair
# break
#if possible: # if all coordinates of the corners are within the image
#counter = counter + 1
yaxis = pair[:, :-1] # y coordinates
xaxis = pair[:, -1:] # x coordinates
# affine transformation to cut the feature area out of the original
# image, all are set to the same size (128,64)
extractedArea = affineTransform(borderReplicate, xaxis[0],yaxis[0],xaxis[1],
yaxis[1],xaxis[2],yaxis[2], M,N)
smoothedExtArea = convolve(extractedArea, gauss(kernel), mode='nearest')
#displayThis(smoothedExtArea)
plt.imshow(smoothedExtArea, cmap=cm.gray)
plt.imsave('./output/saliency3/extfeat2/ara2013_extracted%d_plant%d.png' %(p,imageNum), extractedArea, cmap=cm.gray)
hog = setHoG(smoothedExtArea)
parameters = getParameters(smoothedExtArea, hog)
#showHoG(smoothedExtArea,parameters)
# use p as the id of the feature
parameters = np.insert(parameters, 0, p, 0)
if p == 0:
parametersList = np.array(parameters)
else:
parametersList= np.hstack((parameters, parametersList))
print p
#print parametersList
np.savetxt("./output/saliency3/extfeat2/features_plant%d.csv" %imageNum, parametersList, delimiter=",")
if __name__ == '__main__':
# path to the stored images
mypath = '../Plant_Phenotyping_Datasets/Plant_Phenotyping_Datasets/Plant/Ara2013-Canon/ara2013_rgb_img'
images = getImages(mypath)
for i in range(0,10):
img = images[i]
copy = img # copy to print the img with the saliencyPoints
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # convert to gray
#displayThis(img)
DijKernel = 3 # kernel to get the displacement coordinates
saliencyKernel = 7 # kernel to get the smoothed saliency img
saliency = getSaliency(img, DijKernel, saliencyKernel)
saliencyPoints = getLocalMax(saliency)
N = 128 # hight of the extracted feature image
M = 64 # width of the extracted feature image
featureSmoothing = 3 # kernel to smooth the extracted area
extractFeatures(img, saliencyPoints, featureSmoothing, i, M, N) # use origina image for feature extraction
plt.imshow(saliency,cmap=cm.gray)
plt.imsave('./output/saliency3/ara2013_dk%d_sk%d_plant%d.png' %(DijKernel, saliencyKernel,i), saliency, cmap=cm.gray)
for j in saliencyPoints:
y,x = j.ravel()
cv2.circle(copy, (x,y), 2, 255, -1)
plt.imshow(copy, cmap=cm.gray)
plt.imsave('./output/saliency3/ara2013_dk%d_sk%d_withpoints_plant%d.png' %(DijKernel, saliencyKernel,i), copy, cmap=cm.gray)
#onlyfiles = [f for f in listdir(mypath) if isfile(join(mypath,f))]
#images = np.empty(len(onlyfiles), dtype=object)
#for n in range(0, len(onlyfiles)):
# images[n]= cv2.imread(join(mypath, onlyfiles[n]))
"""
for i in range(0, 10):
#for i in range(0, len(images)):
img = images[i]
#img = cv2.imread('../Plant_Phenotyping_Datasets/Plant_Phenotyping_Datasets/Plant/Ara2013-Canon/ara2013_rgb_img/ara2013_plant001_rgb.png')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
kernel = 3
Dx,Dy = Dij(gray, kernel)
kernel = 7
saliency, blurredSaliency = estimateCenter(Dx,Dy, kernel)
# save images
plt.imshow(blurredSaliency,cmap=cm.Greys)
plt.imsave('./output/saliency/ara2013bs_plant%d.png' %i, blurredSaliency, cmap=cm.Greys)
#inputImg = np.transpose(np.transpose(blurredSaliency)[0])
#print blurredSaliency
#print inputImg
inputImg = blurredSaliency
kernel = 3
Lx = gD(inputImg, kernel, 1, 0)
Ly = gD(inputImg, kernel, 0, 1)
G = sqrt((Lx*Lx)+(Ly*Ly))
pointIndices = getLocalMax(G)
print pointIndices
for j in pointIndices:
y,x = j.ravel()
cv2.circle(inputImg, (x,y), 3, 255, -1)
print inputImg
plt.imshow(blurredSaliency, cmap=cm.Greys)
plt.show()
"""
"""
mypath = './output/saliency'
onlyfiles = [f for f in listdir(mypath) if isfile(join(mypath,f))]
images = np.empty(len(onlyfiles), dtype=object)
for n in range(0, len(onlyfiles)):
images[n]= cv2.imread(join(mypath, onlyfiles[n]))
for i in range(0, 10):
img = images[i]
# transpose to get the first column, transpose back to correct shape
img = np.transpose(np.transpose(img)[0])
#print img
#print img.shape
#print np.transpose(np.transpose(img)[0])
#print np.transpose(np.transpose(img)[0]).shape
#plt.imshow(newimg, cmap=cm.Greys)
#plt.show()
kernel = 3
# gradient in each position
Lx = gD(img, kernel, 1, 0)
Ly = gD(img, kernel, 0, 1)
G = sqrt((Lx*Lx)+(Ly*Ly))
pointIndices = getLocalMax(G)
"""