-
Notifications
You must be signed in to change notification settings - Fork 1
/
index.html
195 lines (184 loc) · 8.47 KB
/
index.html
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
<!DOCTYPE html>
<head>
<meta charset="utf-8">
<title>Photon-Limited Deblurring Dataset</title>
<meta name="viewport" content="width=device-width,initial-scale=1">
<meta name="theme-color" content="#FFFFFF">
<meta name="description" content="Real world dataset for evaluation of deblurring algorithms (both non-blind and blind) in the presence of photon shot noise.">
<meta name="robots" content="index,follow,max-snippet:-1,max-image-preview:large,max-video-preview:-1">
<link rel="icon" type="image/png" href="./img/favicon.ico">
<link rel="apple-touch-icon-precomposed" href="./img/favicon.png">
<!-- styles -->
<link rel="stylesheet" href="./css/bonsai_custom.css">
<link rel="stylesheet" href="./css/style.css">
<!-- fa icons -->
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0-beta3/css/all.min.css">
<!-- code highlighting -->
<link rel="stylesheet" href="./css/a11y-light.min.css">
<script src="./js/highlight.min.js"></script>
<script>hljs.highlightAll();</script>
<!-- interactive elements -->
<script src="./js/magnifier.js"></script>
</head>
<body>
<main style="--maxw:850px; --m:0 auto; --px:20px">
<nav class="bottom_border" style="--py:15px; margin-bottom: 20px;">
<a class="button grey">Home</a>
<a class="button white" href="./pages/tutorial">Tutorial</a>
<a class="button white" target="_blank" href="https://github.com/sanghviyashiitb/poisson-deblurring">Code</a>
</nav>
<div class="container">
<div class="row">
<h1>Photon-Limited Deblurring Dataset</h1>
<h4>Real-world dataset for evaluating deblurring algorithms in photon-limited scenes </h4>
</div>
<div class="flex-container">
<img class="banner-img" src="img/setup1.png">
<div style="width: 5px"></div>
<img class="banner-img" src="img/setup2.png">
</div>
<div class="row">
<h2>Contains</h2>
<ul>
<li>30 low-light photon shot noise corrputed, blurred images in RAW format</li>
<li>Corresponding blur kernel captured using a 30µm pinhole</li>
<li>Ground truth for each image.</li>
</ul>
<div class="center">
<a href="https://1drv.ms/u/s!AjMYTt_aGQ9-hH_myp4irQREzX3K?e=NwARXc" target="_blank" class="button hover-dark black">
<i class="fa fa-download" style="margin-right: 10px" aria-hidden="true"></i>
Download dataset</a>
</div>
<br>
For a description of how to evaluate your deblurring algorithm on this dataset, refer to the file
<a href="https://github.com/sanghviyashiitb/poisson-deblurring/blob/main/demo_synthetic.py"
target="_blank">demo_synthetic.py</a>
in our Github repository.
</div>
<br>
<div class="row">
<h2>Performance</h2>
<h4>Comparing granularity of alternate methods</h4>
<br>
<div class="comparison-container">
<div class="comparison-interactive-container">
<div class="comparison-interactive-img-container">
<img class="comparison-interactive-img" id="img-raw" src="./img/y_23.png">
</div>
<div class="comparison-interactive-meta">
<p>file: y_23.png</p>
<p id="labelX">mouseX: 0</p>
<p id="labelY">mouseY: 0</p>
<p>PSNR: 23.48</p>
<p>SSIM: 0.48</p>
<p>zoomLevel: 4</p>
<p>lightLevel: 0.48</p>
</div>
</div>
<div class="comparison-output-container">
<div class="comparison-output-img-container">
<img class="comparison-output-img" id="dpir_23.png" src="./img/transparent.png">
<p class="center">DPIR</p>
</div>
<div class="comparison-output-img-container">
<img class="comparison-output-img" id="purelet_23.png" src="./img/transparent.png">
<p class="center">Purelet</p>
</div>
<div class="comparison-output-img-container">
<img class="comparison-output-img" id="p4ip_23.png" src="./img/transparent.png">
<p class="center">P4IP</p>
</div>
<div class="comparison-output-img-container">
<img class="comparison-output-img" id="wiener_23.png" src="./img/transparent.png">
<p class="center">Wiener</p>
</div>
</div>
</div>
</div>
<br>
<div class="row">
<h2>Current Benchmarks</h2>
<br>
<table style="--mx: auto">
<thead>
<tr>
<th scope="row">Method</th>
<th>PSNR</th>
<th>SSIM</th>
</tr>
</thead>
<tbody>
<tr>
<td>
<i class="fa fa-crown" style="margin-right: 5px; color: #ffc131" aria-hidden="true"></i>
Unrolled-Poisson PnP [1]</td>
<td>23.48</td>
<td>0.566</td>
</tr>
<tr>
<td>Deep-Wiener Deconvolution [2]</td>
<td>22.85</td>
<td>0.561</td>
</tr>
<tr>
<td>Deep PnP Image Restoration [3]</td>
<td>22.09</td>
<td>0.548</td>
</tr>
<tr>
<td>PURE-LET [4]</td>
<td>20.88</td>
<td>0.501</td>
</tr>
<tr>
<td>RGDN [5]</td>
<td>19.80</td>
<td>0.476</td>
</tr>
</tbody>
</table>
<details>
<summary>References</summary>
<ul style="list-style-type: none;">
<li><cite>[1] Sanghvi, Yash, Abhiram Gnanasambandam, and Stanley H. Chan. "Photon
Limited Non-Blind Deblurring Using Algorithm Unrolling." arXiv
preprint <a href="https://arxiv.org/abs/2110.15314" target="_blank">arXiv:2110.15314 2021</a>
</cite></li>
<li><cite>[2] J. Dong, S. Roth, and B. Schiele, “Deep Wiener deconvolution:
Wiener meets deep learning for image deblurring,” in 34th Conference on Neural
Information Processing Systems, Curran Associates, Inc., 2020
</cite></li>
<li><cite>[3] K. Zhang, W. Zuo, S. Gu, and L. Zhang, “Learning deep CNN denoiser
prior for image restoration,” in Proceedings of the IEEE Conference on Computer
Vision and Pattern Recognition, pp. 3929–3938, 2017.
</cite></li>
<li><cite>[4] J. Li, F. Luisier, and T. Blu, “Pure-let image deconvolution,”
IEEE Transactions on Image Processing, vol. 27, no. 1, pp. 92–105, 2017.
</cite></li>
<li><cite>[5] D. Gong, Z. Zhang, Q. Shi, A. van den Hengel, C. Shen,
and Y. Zhang, “Learning deep gradient descent optimization for
image deconvolution,” IEEE Transactions on Neural Networks and
Learning Systems, vol. 31, no. 12, pp. 5468–5482, 2020
</cite></li>
</details>
</div>
</div>
</main>
<footer>
<div class="footer-content" style="--maxw:850px; width: 100%; --px:20px">
<h6>Site maintained by <a href="https://github.com/aaaakshat" target="_blank">
@aaaakshat</a></h6>
<h6>Dataset maintained by <a href="https://github.com/sanghviyashiitb" target="_blank">
@sanghviyashiitb</a></h6>
<br>
<h6>Brought to you by the <a href="https://engineering.purdue.edu/ChanGroup/" target="_blank">i2Lab</a>
at Purdue University.</h6>
</div>
</footer>
<script>
/* Initiate Magnify Function
with the id of the image, and the strength of the magnifier glass:*/
magnify("img-raw", "dpir_23.png", "purelet_23.png", "p4ip_23.png", "wiener_23.png", 4);
</script>
</body>
</html>