-
Notifications
You must be signed in to change notification settings - Fork 8
/
CircularDofFilterGen.py
executable file
·208 lines (183 loc) · 9.12 KB
/
CircularDofFilterGen.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
200
201
202
203
204
205
206
207
208
#!/usr/bin/python
import argparse
import math
from functools import reduce
"""
********************************************************************
********************************************************************
* Generated Filter by CircularDofFilterGenerator tool *
* Copyright (c) Kleber A Garcia ([email protected])*
* https://github.com/kecho/CircularDofFilterGenerator *
********************************************************************
********************************************************************
**")
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
"""
def generateFilter(lan, r, c, t):
if t <= -1:
print("Invalid transition bandwidth. Must be greater than -1 and preferably positive.");
return;
P = []
## a b A B
a = 0
b = 1
A = 2
B = 3
if c == 1:
P = [[1.624835, -0.862325, 0.767583, 1.862321]]
elif c == 2:
P = [
[ 5.268909, -0.886528, 0.411259, -0.548794 ],
[ 1.558213, -1.960518, 0.513282, 4.561110 ]]
elif c == 3:
P = [
[ 5.043495, -2.176490, 1.621035, -2.105439 ],
[ 9.027613, -1.019306, -0.280860, -0.162882 ],
[ 1.597273, -2.815110, -0.366471, 10.300301 ]]
elif c == 4:
P = [
[ 1.553635, -4.338459, -5.767909, 46.164397],
[ 4.693183, -3.839993, 9.795391, 15.227561 ],
[ 8.178137, -2.791880, -3.048324, 0.302959 ],
[ 12.328289, -1.342190, 0.010001, 0.244650 ]]
elif c == 5:
P = [
[ 1.685979, -4.892608, -22.356787, 85.912460 ],
[ 4.998496, -4.711870, 35.918936 , -28.875618 ],
[ 8.244168, -4.052795, -13.212253, -1.578428 ],
[ 11.900859, -2.929212, 0.507991, 1.816328 ],
[ 16.116382, -1.512961, 0.138051, -0.010000 ]]
else:
print("Invalid component count. Must be [1-5].");
return;
def KernelFun(x, C):
return (
math.cos(x*x*C[a]) * math.exp( x * x * C[b]), #real
math.sin(x*x*C[a]) * math.exp( x * x * C[b]), #imaginary
C[A], #real weight
C[B] #imaginary weight
)
totalBandwidth = 1.0 + t
kernels = [[KernelFun(totalBandwidth * float(i)/float(r), C) for i in range(-r,r+1,1)] for C in P]
#normalize kernels
accum = 0.0
for k in kernels:
for v in k:
for w in k:
accum = accum + v[A]*(v[0]*w[0] - v[1]*w[1]) + v[B]*(v[0]*w[1] + v[1]*w[0])
normConstant = 1.0 / math.sqrt(accum)
kernelsNormalized = [[ (normConstant*real, normConstant*im, 0.0, 0.0) for (real, im, Av, Bv) in k ] for k in kernels]
#bracket the kernel so we maximize precision. This means figureout a Offset and a Scale
# real imaginary
scales = []
offsets = [reduce((lambda v1, v2: (min(v1[0],v2[0]),min(v1[1],v2[1]))), k) for k in kernelsNormalized]
for (k,o) in zip(kernelsNormalized, offsets):
scale = (0.0 ,0.0)
for v in k:
realScale = v[0] - o[0]
immScale = v[1] - o[1]
scale = (scale[0]+realScale, scale[1]+immScale)
scales.append(scale)
#print(offsets)
#print(scales)
finalKernels = [[(v[0],v[1],(v[0]-o[0])/s[0],(v[1]-o[1])/s[1]) for v in k] for (k,o,s) in zip(kernelsNormalized, offsets, scales)]
componentWeights = [ (comp[2], comp[3]) for comp in P ]
if lan=="hlsl":
printHlsl(r, finalKernels, componentWeights, offsets, scales)
else:
printGlsl(r, finalKernels, componentWeights, offsets, scales)
# Visualize the kernel
shouldVisualize = False # Set to True to visuzlize
saveToFile = "kernel.png" # or None
if not shouldVisualize: return
try:
from PIL import Image, ImageFilter
except ImportError:
print("/** Pillow not found! Try running [python3 -m pip install --upgrade Pillow] if you want to visualize the kernel **/")
return
d = r * 2 + 1
realKernel = [[0 for x in range(d)] for y in range(d)]
# Note that the kernels are already the result of a horizontal blur of a single white pixel (val = 1.0)
# To visualize it we only need a vertical blur
for x in range(d):
for y in range(d):
for i in range(len(finalKernels)):
k = finalKernels[i]
A, B = componentWeights[i][0], componentWeights[i][1]
Pr, Pi = k[x][0], k[x][1]
Qr, Qi = k[y][0], k[y][1]
R, I = Pr * Qr - Pi * Qi, Pr * Qi + Pi * Qr
S = A * R + B * I
realKernel[x][y] = realKernel[x][y] + S
maxV = 0
for x in range(d):
for y in range(d):
maxV = max(maxV, abs(realKernel[x][y]))
maxV = 128 / maxV
pixelW = 8
width = d * pixelW
height = d * pixelW
img = Image.new("RGB", (width, height))
imgMap = img.load()
for y in range(height):
for x in range(width):
val = 128 + (int)(maxV * realKernel[(int)(x/pixelW)][(int)(y/pixelW)])
imgMap[x, y] = (val, val, val)
if (saveToFile):
img.save(saveToFile)
img.show()
def printHlsl(r, finalKernels, componentWeights, offsets, scales):
syntax = ("uint", "float", "static const", "{", "};")
printShaderCommon(r, finalKernels, componentWeights, offsets, scales, syntax)
def printGlsl(r, finalKernels, componentWeights, offsets, scales):
syntax = ("int", "vec", "const", "vec4[](", ");")
printShaderCommon(r, finalKernels, componentWeights, offsets, scales, syntax)
def printShaderCommon(r, finalKernels, componentWeights, offsets, scales, syntax):
print("/********************************************************************/")
print("/********************************************************************/")
print("/* Generated Filter by CircularDofFilterGenerator tool */")
print("/* Copyright (c) Kleber A Garcia ([email protected])*/")
print("/* https://github.com/kecho/CircularDofFilterGenerator */")
print("/********************************************************************/")
print("/********************************************************************/")
print("/**")
print(""" THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE. """);
print("**/")
diameter = r * 2 + 1;
print("%s %s KERNEL_RADIUS = %d;" % (syntax[2], syntax[0], r))
print("%s %s KERNEL_COUNT = %d;" % (syntax[2], syntax[0], diameter))
filterCount = len(finalKernels);
for i in range(0,filterCount):
k = finalKernels[i]
o = offsets[i]
s = scales[i]
comp = componentWeights[i]
print("%s %s4 Kernel%dBracketsRealXY_ImZW = %s4(%f,%f,%f,%f);" % (syntax[2], syntax[1],i, syntax[1], o[0],s[0],o[1],s[1]) )
print("%s %s2 Kernel%dWeights_RealX_ImY = %s2(%f,%f);" % (syntax[2],syntax[1], i, syntax[1], comp[0],comp[1]) )
print("%s %s4 Kernel%d_RealX_ImY_RealZ_ImW[] = %s" % (syntax[2],syntax[1], i, syntax[3]))
for pixel in range(0,diameter):
val = k[pixel]
print("\t%s4(/*XY: Non Bracketed*/%f,%f,/*Bracketed WZ:*/%f,%f)%s" % (syntax[1], val[0],val[1],val[2],val[3], "," if pixel < (diameter-1) else ""))
print("%s" % syntax[4])
def main():
parser = argparse.ArgumentParser('Circular Dof Filter Generator. Kleber Garcia (c) 2017.\n\nPublication: http://dl.acm.org/citation.cfm?id=3085022.\nShader toy example: https://www.shadertoy.com/view/Xd2BWc\n')
parser.add_argument('-l', dest='Language', metavar='Language', type=str, help='Language to use. Default is hlsl, possible values "hlsl" or "glsl".', choices=["hlsl","glsl"], default="hlsl");
parser.add_argument('-r', dest='FilterRadius', metavar='FilerRadius', type=int, help='Filter Radius (in pixels). Default is 8 (diameter of 17)', default=8);
parser.add_argument('-c', dest='Components', metavar='ComponentCount', type=int, help='Component count. Default is 2.', default=2);
parser.add_argument('-t', dest='Transition', metavar='TransitionBandwidth', type=float, help='Transition bandwidth. Default is 0. Values of [0.2 .. 0.4] work best.', default=0);
args = parser.parse_args()
generateFilter(args.Language, args.FilterRadius, args.Components, args.Transition)
if __name__ == "__main__":
main()