-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
263 lines (246 loc) · 12.4 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
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
<!DOCTYPE html
PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>NTIRE 2023: 360° Super-Resolution Challenge</title>
<!--=================Meta tags==========================-->
<meta name="robots" content="index,follow">
<meta name="description"
content="Deep convolutional neural network has demonstrated its capability of learning a deterministic mapping for the desired imagery effect. However, the large variety of user flavors motivates the possibility of continuous transition among different output effects. Unlike existing methods that require a specific design to achieve one particular transition (e.g., style transfer), we propose a simple yet universal approach to attain a smooth control of diverse imagery effects in many low-level vision tasks, including image restoration, image-to-image translation, and style transfer. Specifically, our method, namely Deep Network Interpolation (DNI), applies linear interpolation in the parameter space of two or more correlated networks. A smooth control of imagery effects can be achieved by tweaking the interpolation coefficients. In addition to DNI and its broad applications, we also investigate the mechanism of network interpolation from the perspective of learned filters.">
<meta name="keywords" content="interpolation, super resolution, style transfer, cyclegan">
<link rel="author" href="https://xinntao.github.io/projects/DNI">
<!--=================js==========================-->
<link href="./css.css" rel="stylesheet" type="text/css">
<link rel="stylesheet" type="text/css" href="./project.css" media="screen">
<script src="./effect.js "></script>
<!-- Latex -->
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
tex2jax: {inlineMath: [['$','$'], ['\\(','\\)']]},
TeX: { equationNumbers: { autoNumber: "AMS" } },
});
</script>
<script type="text/javascript" async
src="js/MathJax.js">
</script>
<!--=================Google Analytics==========================-->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-129775907-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'UA-129775907-1');
</script>
</head>
<style type="text/css">
.tg {border-collapse:collapse;border-spacing:0;}
.tg td{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
overflow:hidden;padding:10px 5px;word-break:normal;}
.tg th{border-color:black;border-style:solid;border-width:1px;font-family:Arial, sans-serif;font-size:14px;
font-weight:normal;overflow:hidden;padding:10px 5px;word-break:normal;}
.tg .tg-cly1{text-align:left;vertical-align:middle}
.tg .tg-fp3b{border-color:#000000;font-weight:bold;text-align:left;vertical-align:middle}
.tg .tg-yla0{font-weight:bold;text-align:left;vertical-align:middle}
</style>
<body>
<div id="content">
<div id="content-inner">
<div class="section head">
<h1>
NTIRE 2023: <font color="Tomato">360</font>° <font color="Tomato">S</font>uper-<font color="Tomato">R</font>esolution Challenge Dataset
</h1>
<!--=================Authors==========================-->
<!-- <div class="authors">
<a href="https://xinntao.github.io/" target="_blank">Xintao Wang</a> <sup>1</sup>
<a href="https://yuke93.github.io/" target="_blank">Ke Yu</a> <sup>1</sup>
<a href=" https://scholar.google.com.hk/citations?user=OSDCB0UAAAAJ" target="_blank">Chao Dong</a>
<sup>2</sup>
<a href="https://scholar.google.com/citations?user=qpBtpGsAAAAJ" target="_blank">Xiaoou Tang</a> <sup>1</sup>
<a href="http://personal.ie.cuhk.edu.hk/~ccloy/" target="_blank">Chen Change Loy</a> <sup>3</sup>
</div>
<div class="affiliations ">
<sup>1</sup> <a href="http://mmlab.ie.cuhk.edu.hk/" target="_blank">CUHK - SenseTime Joint Lab, The Chinese
University of Hong Kong</a> <br>
<sup>2</sup> Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences<br>
<sup>3</sup> Nanyang Technological University, Singapore
</div>-->
<!--=================Tabs==========================-->
<ul id="tabs">
<li><a href="#Flickr360" name="#tab2">Flickr360</a></li>
<li><a href="#ODV360" name="#tab3">ODV360</a></li>
</div>
<br>
<!--=================Abstract==========================-->
<div class="section abstract">
<center>
<div class="rightView">
<img class="small" src="./src/mark.png" width="400">
</div>
</center>
<h3>Please see more competition information in <a href="https://github.com/360SR/360SR-Challenge", target="_blank">https://github.com/360SR/360SR-Challenge</a>.</h3>
<h2>Abstract</h2>
<br>
<p>
The 360° or omnidirectional images/videos can provide an immersive and interactive experience and have received
much research attention with the popularity of AR/VR applications. Unlike planar images/videos that have a narrow
field of view (FoV), 360° images/videos can represent the whole scene in all directions. However, 360° images/videos
suffer from the lower angular resolution problem since they are captured by fisheye lens with the same sensor size
for capturing planar images. Although the whole 360° images/videos are of high resolution, the details are not
satisfying. In many application scenarios, increasing the resolution of 360° images/videos is highly demanded
to achieve higher perceptual quality and boost the performance of downstream tasks.
</p>
<p>
Recently, considerable success has been achieved in the image and video super-resolution (SR) task with the
development of deep learning-based methods. Although 360° images/videos are often transformed into 2D planar
representations by preserving omnidirectional information in practice, like equirectangular projection (ERP)
and cube map projection (CMP), existing super-resolution methods still cannot be directly applied to 360°
images/videos due to the distortions introduced by the projections. As for videos, the temporal relationships
in a 360° video should be further considered since it is different from that in an ordinary 2D video.
Therefore, how to effectively super-resolve 360° image/video by considering these characteristics remains challenging.
</p>
<p>
To rectify the lack of high-quality datasets in the community of omnidirectional image/video super-resolution, we construct new 360° datasets for image (<a href="#Flickr360">Flickr360</a>) and video (<a href="#ODV360">ODV360</a>), respectively.
</p>
</div>
<!--=================Flickr360==========================-->
<br>
<div class="section" , id="Flickr360">
<h2>Flickr360</h2>
<center>
<div class="rightView">
<img class="small" src="./src/flickr360.png", width="100%">
</div>
</center>
<p>
Flickr360 contains about 3150 ERP images with a resolution larger than 5k. Specifically,
3100 images are collected from Flickr, and the other 50 images are captured by Insta360°
cameras. The images from Flickr are under either Creative Commons BY 2.0, Creative Commons BY-NC 2.0,
Public Domain Mark 1.0, Public Domain CC0 1.0, or U.S. Government Works license. All of these licenses
allow free use, redistribution, and adaptation for non-commercial purposes. The image contents vary both
indoors and outdoors. We first downsample the original images into 2k resolution (2048 x 1024), serving
as HR images. These HR images are further downsampled into LR images. The data partition is shown in the
following table.
</p>
<p>
<center>
<table class="tg">
<thead>
<tr>
<th></th>
<th>Training</th>
<th>Validation</th>
<th>Testing</th>
</tr>
</thead>
<tbody>
<tr>
<td>Source</td>
<td>Flickr 360</td>
<td>Flickr 360</td>
<td>Flickr 360+capturing</td>
</tr>
<tr>
<td>Number</td>
<td>3000</td>
<td>50</td>
<td>50+50</td>
</tr>
<tr>
<td>Storage</td>
<td>8.1G (HR) + 553M (LR)</td>
<td>137M (HR) + 9.3M (LR)</td>
<td>271M (HR) + 20M (LR)</td>
</tr>
</tbody>
</table>
</center>
</p>
<h3>
Download
</h3>
<p>● <a href="https://drive.google.com/drive/folders/1lDIxTahDXQ5w5x_UZySX2NOes_ZoNztN", target="_blank">Google Drive</a></p>
<p>● <a href="https://share.weiyun.com/6p2fsaxO", target="_blank">腾讯微云</a></p>
<h3>
Samples
</h3>
<p>See <a href="./Flickr360/Flickr360.html", target="_blank">here</a> for samples. </p>
</div>
<br>
<!--=================ODV360==========================-->
<div class="section" , id="ODV360">
<h2>ODV360</h2>
<center>
<div class="rightView">
<img class="small" src="./src/odv360.png", width="100%">
</div>
</center>
<p>
ODV360 including two parts:
</p><p>
● 90 videos collected from YouTube and public 360° video dataset
</p><p>
These videos are carefully selected and have high quality to be used for restoration.
All videos have the license of Creative Commons Attribution license (reuse allowed), and our dataset is used only for academic and research proposes
</p><p>
● 160 videos collected by ourselves with Insta360 cameras
</p><p>
The cameras we use include Insta 360 X2 and Insta 360 ONE RS. They can capture high-resolution (5.7K) omnidirectional videos.
</p><p>
These collected omnidirectional videos cover a large range of diversity, and the video contents vary indoors and outdoors.
To facilitate the use of these videos for research, we downsample the original videos into 2K resolution (2160x1080) by OpenCV.
The number of frames per video is fixed at about 100. We randomly divide these videos into training, validation, and testing sets,
as shown in the following table.
</p>
<p>
<center>
<table class="tg">
<thead>
<tr>
<th></th>
<th>Training</th>
<th>Validation</th>
<th>Testing</th>
<th>All</th>
</tr>
</thead>
<tbody>
<tr>
<td>Number</td>
<td>210</td>
<td>20</td>
<td>20</td>
<td>250</td>
</tr>
<tr>
<td>Storage</td>
<td>GT(59G)+LR(4.9G)</td>
<td>GT(5.3G)+LR(446M)</td>
<td>GT(5.7G)+LR(485M)</td>
<td>75.8G</td>
</tr>
</tbody>
</table>
</center>
</p>
<h3>
Download
</h3>
<p>● <a href="https://drive.google.com/drive/folders/1lDIxTahDXQ5w5x_UZySX2NOes_ZoNztN", target="_blank">Google Drive</a></p>
<p>● <a href="https://share.weiyun.com/8zd7X0TZ", target="_blank">腾讯微云</a></p>
<h3>
Samples
</h3>
<p>See <a href="./odv360_test/ODV360.html", target="_blank">here</a> for samples. </p>
</div>
<br>
<!--=================Contact==========================-->
<div class="section contact">
<h2 id="contact">Contact</h2>
<p>If you have any question, please open an issue in <a href="https://github.com/360SR/360SR-Challenge", target="_blank">https://github.com/360SR/360SR-Challenge</a>.</p>
</div>
</body>
</html>