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
| #! /usr/bin/python
from PIL import Image
import numpy
from fractions import gcd
from functools import reduce
import random
import sys
# file meta data
__version__ = "1.0"
__author__ = "abgoyal"
__license__ = "apl"
## some utilities
#sort a 2d list by a given column #
rev_sortbycolumn = lambda l,c: sorted(l, key = lambda x: x[c], reverse=True)
sortbycolumn = lambda l,c: sorted(l, key = lambda x: x[c])
# vector gcd
vgcd = lambda v: reduce(gcd,v)
## algo implementations
# distance metric - this it the key
def distance(v1, v2): # fractional difference in pixel values
return numpy.sum(((v1-v2)/(v1+v2+1))**2)
# simple cluster detection
def cluster1d(values):
z = sortbycolumn(values, 1)
part = [ z.pop()[0]]
while vgcd(part)>1:
part.append(z.pop()[0])
return vgcd(part[:-1])
# estimate the strip width
def stripwidth_algo(im):
#extract image pixel data as a numpy array
image_data = numpy.asarray(im.convert("L").getdata(), 'float')
image_width = im.size[0]
image_height = im.size[1]
image_column = lambda i: (image_data[i::image_width])
# calculate correlation coefs between each pair of adjacent columns in the image
dists = [ [i+1, distance(image_column(i), image_column(i+1))] for i in xrange(0,image_width-1) ]
return cluster1d(dists)
# unshred an image given strip width
def unshred_algo(im,strip_width):
im_width= im.size[0]
im_height = im.size[1]
n = im_width/strip_width # number of strips
# extract grayscale pixel data from image into a numpy.array
image_data = numpy.asarray(im.convert("L").getdata(), 'float')
# extracts all pixels from a specific column
image_column = lambda i: (image_data[i::im_width])
# extract the left and right edges of each strip as columns
left_edges = []
right_edges = []
for i in xrange(0,n):
left_edges.append(image_column(i*strip_width))
right_edges.append(image_column(i*strip_width + strip_width -1))
# calculate the distance metric between each left-edge<>right-edge pair
dists = []
for left_edge_index,left_edge in enumerate(left_edges):
for right_edge_index,right_edge in enumerate(right_edges):
if not left_edge_index == right_edge_index: # a strip cannot be next to itself
dists.append([left_edge_index, right_edge_index, distance(left_edge,right_edge) ])
sorted_dists = rev_sortbycolumn(dists,2)
# select the pairs with the least distance between (strip_left_edge, strip_right_edge) pixels
assigned_left_edges = []
assigned_right_edges = {}
while len(assigned_right_edges)<n:
left_edge_index, right_edge_index, dist = sorted_dists.pop()
if (right_edge_index not in assigned_right_edges) and (left_edge_index not in assigned_left_edges):
assigned_right_edges[right_edge_index] = left_edge_index
assigned_left_edges.append(left_edge_index)
# last strip assigned is most probably the left edge
s = assigned_left_edges[-1]
order = [ s ]
while s in assigned_right_edges and len(order)<n:
order.append(assigned_right_edges[s])
s = assigned_right_edges[s]
return order
def strip_copy(im_original, im_copy, w, a, b):
h = im_original.size[1]
a_tuple= tuple([a*w, 0, (a+1)*w, h])
b_tuple= tuple([b*w, 0, (b+1)*w, h])
im_copy.paste( im_original.crop(a_tuple) , box=b_tuple )
def unshred(im,strip_width=None):
im2 = im.copy()
# if strip width not given, try to calculate it
if not strip_width:
strip_width = stripwidth_algo(im)
order = unshred_algo(im,strip_width)
for i,j in enumerate(order):
strip_copy(im, im2, strip_width, j, i)
return im2
def shred(im,m):
im_width= im.size[0]
im_height = im.size[1]
im2 = im.copy()
n = im_width/m
strips = range(0,n)
random.shuffle(strips)
for f,t in enumerate(strips):
strip_copy(im, im2, m, f, t)
return im2
def demo():
im_shreded = Image.open("TokyoPanoramaShredded.png")
im_fixed1 = unshred(im_shreded)
im_fixed1.show()
im_fixed2 = unshred(im_shreded,32)
im_fixed2.show()
im_reshredded = shred(im_fixed2, 16)
im_fixed1 = unshred(im_shreded)
im_fixed1.show()
im_fixed2 = unshred(im_shreded,16)
im_fixed2.show()
def process_command(args):
cmd = args[0]
if cmd == "shred":
w = int(args[1])
ifname = args[2]
ofname = args[3]
im = Image.open(ifname)
im_shredded = shred(im, w)
im_shredded.save(ofname)
return
if cmd == "unshred":
w = int(args[1])
ifname = args[2]
ofname = args[3]
im = Image.open(ifname)
if w==0:
im_fixed = unshred(im)
else:
im_fixed = unshred(im,w)
im_fixed.save(ofname)
return
print "Unknown command"
return
if __name__ == "__main__":
if len(sys.argv) > 1:
process_command(sys.argv[1:])
else:
demo()
|
