-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest-restir-spatial-spatial.py
111 lines (88 loc) · 3.24 KB
/
test-restir-spatial-spatial.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
import mitsuba as mi
import drjit as dr
import matplotlib.pyplot as plt
from tqdm import tqdm
mi.set_variant("cuda_ad_rgb")
import restirgi
if __name__ == "__main__":
n_iterations = 100
spp = 1
scene = mi.cornell_box()
scene["sensor"]["film"]["width"] = 1024
scene["sensor"]["film"]["height"] = 1024
scene["sensor"]["film"]["rfilter"] = mi.load_dict({"type": "box"})
scene: mi.Scene = mi.load_dict(scene)
# scene = mi.load_file("./data/scenes/wall/scene.xml")
scene = mi.load_file("./data/scenes/staircase/scene.xml")
# scene: mi.Scene = mi.load_file("data/scenes/shadow-mask/scene.xml")
ref = mi.render(scene, spp=256)
mi.util.write_bitmap("out/ref.exr", ref)
biased: restirgi.RestirIntegrator = mi.load_dict(
{
"type": "restirgi",
"jacobian": False,
"bias_correction": False,
"spatial_spatial_reuse": True,
"bsdf_sampling": True,
"max_M_spatial": 500,
"max_M_temporal": 30,
}
)
unbiased: restirgi.RestirIntegrator = mi.load_dict(
{
"type": "restirgi",
"jacobian": False,
"bias_correction": False,
"spatial_spatial_reuse": False,
"bsdf_sampling": True,
"max_M_spatial": 500,
"max_M_temporal": 30,
}
)
var_biased = []
bias_biased = []
mse_biased = []
print("Biased")
for i in tqdm(range(n_iterations)):
img = mi.render(scene, integrator=biased, seed=i, spp=spp)
var_biased.append(dr.mean_nested(dr.sqr(img - dr.mean_nested(img)))[0])
bias_biased.append(dr.mean_nested(img - ref)[0])
mse_biased.append(dr.mean_nested(dr.sqr(img - ref)))
img_biased = img
mi.util.write_bitmap("out/biased.exr", img_biased)
var_unbiased = []
bias_unbiased = []
mse_unbiased = []
print("Unbiased")
for i in tqdm(range(n_iterations)):
img = mi.render(scene, integrator=unbiased, seed=i, spp=spp)
var_unbiased.append(dr.mean_nested(dr.sqr(img - dr.mean_nested(img)))[0])
bias_unbiased.append(dr.mean_nested(img - ref)[0])
mse_unbiased.append(dr.mean_nested(dr.sqr(img - ref)))
img_unbiased = img
mi.util.write_bitmap("out/unbiased.exr", img_unbiased)
fig, ax = plt.subplots(2, 3, figsize=(20, 10))
fig.patch.set_visible(False)
ax[0][0].axis("off")
ax[0][0].imshow(mi.util.convert_to_bitmap(ref))
ax[0][0].set_title("Reference")
ax[0][1].plot(bias_biased, label="Biased")
ax[0][1].plot(bias_unbiased, label="Bias Corrected")
ax[0][1].legend(loc="best")
ax[0][1].set_title("Sample Bias")
ax[1][0].axis("off")
ax[1][0].set_title("Biased")
ax[1][0].imshow(mi.util.convert_to_bitmap(img_biased))
ax[1][1].axis("off")
ax[1][1].set_title("Bias Corrected")
ax[1][1].imshow(mi.util.convert_to_bitmap(img_unbiased))
ax[0][2].plot(mse_biased, label="Biased")
ax[0][2].plot(mse_unbiased, label="Bias Corrected")
ax[0][2].legend(loc="best")
ax[0][2].set_title("MSE")
ax[1][2].plot(var_biased, label="Biased")
ax[1][2].plot(var_unbiased, label="Bias Corrected")
ax[1][2].legend(loc="best")
ax[1][2].set_title("Variance")
fig.tight_layout()
plt.show()