-
Notifications
You must be signed in to change notification settings - Fork 1
/
DirectionNet.html
173 lines (146 loc) · 7.52 KB
/
DirectionNet.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
<!DOCTYPE html>
<html><head lang="en"><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<meta http-equiv="x-ua-compatible" content="ie=edge">
<title>Wide-Baseline Relative Camera Pose Estimation with Directional Learning</title>
<meta name="description" content="">
<meta name="viewport" content="width=device-width, initial-scale=1">
<!-- <base href="/"> -->
<!-- <link rel="apple-touch-icon" href="apple-touch-icon.png"> -->
<!-- <link rel="icon" type="image/png" href=""> -->
<!-- Place favicon.ico in the root directory -->
<link rel="stylesheet" href="./DirectionNet/bootstrap.min.css">
<link rel="stylesheet" href="./DirectionNet/font-awesome.min.css">
<link rel="stylesheet" href="./DirectionNet/codemirror.min.css">
<link rel="stylesheet" href="./DirectionNet/app.css">
<link rel="stylesheet" href="./DirectionNet/bootstrap.min(1).css">
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-E0ZMW34H4P"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-E0ZMW34H4P');
</script>
<script src="./DirectionNet/jquery.min.js"></script>
<script src="./DirectionNet/bootstrap.min.js"></script>
<script src="./DirectionNet/codemirror.min.js"></script>
<script src="./DirectionNet/clipboard.min.js"></script>
<script src="./DirectionNet/app.js"></script>
</head>
<body>
<div class="container" id="main">
<div class="row">
<h1 class="col-md-12 text-center">
Wide-Baseline Relative Camera Pose Estimation with Directional Learning<br>
<small>
CVPR 2021
</small>
</h1>
</div>
<div class="row">
<div class="col-md-12 text-center">
<ul class="list-inline">
<li>
<a href="https://arthurchen0518.github.io/">
Kefan Chen
</a>
</li>
<li>
<a href="https://www.cs.cornell.edu/~snavely/">
Noah Snavely
</a>
</li>
<li>
<a href="http://www.ameeshmakadia.com/index.html">
Ameesh Makadia
</a>
</li>
</ul>
Google Research <br>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2 text-center">
<ul class="nav nav-pills nav-justified">
<li>
<a href="https://arxiv.org/abs/2106.03336">
<img src="./DirectionNet/paper.PNG" height="120px"><br>
<h4><strong>Paper</strong></h4>
</a>
</li>
<li>
<a href="https://youtu.be/tJ2JTFmi0oI">
<img src="./DirectionNet/youtube_icon_dark.png" height="120px"><br>
<h4><strong>Technical Video </strong></h4>
</a>
</li>
<li>
<a href="https://github.com/arthurchen0518/DirectionNet">
<img src="./DirectionNet/github_pad.png" height="120px"><br>
<h4><strong>Code</strong></h4>
</a>
</li>
</ul>
</div>
<br>
</div>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Example Real Results: View Synthesis, Relighting and Material editing
</h3>
<video id="v0" width="100%" autoplay="" loop="" muted="" controls="">
<source src="files/real-world_results.mp4" type="video/mp4">
</video>
</div>
</div> -->
<div class="row">
<div class="col-md-8 col-md-offset-2">
<img src="./DirectionNet/TrainingPipeline.png" class="img-responsive" alt="overview"><br>
<h3>
Abstract
</h3>
<p class="text-justify">
Modern deep learning techniques that regress the relative camera pose between two images have difficulty dealing with challenging scenarios, such as large camera motions resulting in occlusions and significant changes in perspective that leave little overlap between images. These models continue to struggle even with the benefit of large supervised training datasets. To address the limitations of these models, we take inspiration from techniques that show regressing keypoint locations in 2D and 3D can be improved by estimating a discrete distribution over keypoint locations. Analogously, in this paper we explore improving camera pose regression by instead predicting a discrete distribution over camera poses. To realize this idea, we introduce DirectionNet, which estimates discrete distributions over the 5D relative pose space using a novel parameterization to make the estimation problem tractable. Specifically, DirectionNet factorizes relative camera pose, specified by a 3D rotation and a translation direction, into a set of 3D direction vectors. Since 3D directions can be identified with points on the sphere, DirectionNet estimates discrete distributions on the sphere as its output. We evaluate our model on challenging synthetic and real pose estimation datasets constructed from Matterport3D and InteriorNet. Promising results show a near 50% reduction in error over direct regression methods.
</p>
</div>
</div>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Technical Video
</h3>
<div class="text-center">
<div style="position:relative;padding-top:56.25%;">
<iframe src="./files/...html" allowfullscreen="" style="position:absolute;top:0;left:0;width:100%;height:100%;"></iframe>
</div>
</div>
</div>
</div> -->
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
BibTeX
</h3>
<div class="form-group col-md-10 col-md-offset-1">
<textarea id="bibtex" class="form-control" readonly="" style="display: none;">
@InProceedings{Chen_2021_CVPR,
author = {Chen, Kefan and Snavely, Noah and Makadia, Ameesh},
title = {Wide-Baseline Relative Camera Pose Estimation With Directional Learning},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2021},
pages = {3258-3268}
}
</textarea>
</div>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
The website template was borrowed from <a href="http://mgharbi.com/">Michaël Gharbi</a>.
<p></p>
</div>
</div>
<!-- </div> -->
</body></html>