-
Notifications
You must be signed in to change notification settings - Fork 1
/
spot.html
224 lines (220 loc) · 9.21 KB
/
spot.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
<!DOCTYPE HTML>
<!-- THE DOCTYPE tells the browser what type of webpage to render, the doctype used in this website is HTML5 which is the most recent HTML. -->
<html lang="en-US">
<head>
<meta charset="UTF-8">
<title>SPOT: Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints</title>
<meta name="description" content="Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints.">
<style>
.title{
font-family: Arial, Helvetica, sans-serif;
text-align:center;
margin-top:30pt;
margin-bottom:0pt;
margin-left: auto;
margin-right: auto;
color:#303030;
font-size:33pt;
font-weight:800;
letter-spacing:0pt;
max-width:1200px;
}
.subtitle{
font-family: Arial, Helvetica, sans-serif;
text-align:center;
margin-top:8pt;
margin-bottom:0;
margin-left: auto;
margin-right: auto;
max-width:1100px;
color:#303030;
font-size:30pt;
font-weight:500;
letter-spacing:0pt;
}
.section_title{
font-family: Arial, Helvetica, sans-serif;
text-align:left;
padding-left:25pt;
padding-right:25pt;
color:#181818;
font-size:25pt;
font-weight:500;
letter-spacing:0pt;
}
.subsection_title{
font-family: Arial, Helvetica, sans-serif;
text-align:left;
padding-left:25pt;
padding-right:25pt;
padding-top:25pt;
color:#504d4d;
font-size:20;
font-weight:500;
letter-spacing:0pt;
}
.body{
font-family: Arial, Helvetica, sans-serif;
color:#000000;
padding-right:25pt;
padding-left:25pt;
padding-bottom:25pt;
text-align:justify;
font-size:12pt;
font-weight:300;
letter-spacing:0pt;
line-height: 20pt;
}
.col{
font-family: Arial, Helvetica, sans-serif;
text-align: center;
float: left;
width: 16%;
color:#000000;
font-size:15pt;
font-weight:100;
letter-spacing:0pt;
height: 25pt;
}
.row{
margin-left: auto;
margin-right: auto;
max-width:1200px;
}
.row:after {
content: "";
display: table;
clear: both;
}
.affiliation{
font-family: Arial, Helvetica, sans-serif;
text-align: center;
color:#000000;
font-size:13pt;
font-weight:100;
letter-spacing:0pt;
}
.contribution{
font-family: Arial, Helvetica, sans-serif;
text-align: center;
color:#000000;
font-size:10pt;
font-weight:100;
letter-spacing:0pt;
}
.publish{
font-family: Arial, Helvetica, sans-serif;
text-align: center;
color:#000000;
font-size:20pt;
font-weight:600;
letter-spacing:0pt;
}
.container{
margin-left:auto;
margin-right:auto;
margin-top: 5pt;
max-width:1200px
}
</style>
</head>
<body style="background-color:#ffffff;">
<div class="section">
<h1 class="title">SPOT: Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints</h1>
<!-- Author name-->
<div class="row">
<div class="col" style="font-weight: 200; height: auto; margin-bottom: -22pt; width: 25%;">
<p>Spencer Carmichael<sup>1</sup></p>
</div>
<div class="col" style="font-weight: 200; height: auto; margin-bottom: -22pt; width: 25%;">
<p>Rahul Agrawal<sup>1*</sup></p>
</div>
<div class="col" style="font-weight: 200; height: auto; margin-bottom: -22pt; width: 25%;">
<p>Ram Vasudevan<sup>1,2</sup></p>
</div>
<div class="col" style="font-weight: 200; height: auto; margin-bottom: -22pt; width: 25%;">
<p>Katherine A. Skinner<sup>1</sup></p>
</div>
</div>
<!-- Mailbox -->
<div class="row">
<div class="col", style="font-size:10pt; width: 25%;">
<p>[email protected]</p>
</div>
<div class="col", style="font-size:10pt; width: 25%;">
<p>[email protected]</p>
</div>
<div class="col", style="font-size:10pt; width: 25%;">
<p>[email protected]</p>
</div>
<div class="col", style="font-size:10pt; width: 25%;">
<p>[email protected]</p>
</div>
</div>
<!-- Affiliation -->
<div class="contribution">
<p><sup>1</sup>Robotics Department, University of Michigan, Ann Arbor</p>
</div>
<div class="contribution">
<p><sup>2</sup>Department of Mechanical Engineering, University of Michigan, Ann Arbor</p>
</div>
<div class="contribution">
<p><sup>*</sup>Rahul Agrawal contributed to this work while employed at University of Michigan.</p>
</div>
<!-- Links -->
<div class="row" style="margin-left:auto; margin-right:auto; margin-top:10pt;max-width:800px;">
<div class="col" style="width: 33%; height: 60pt;">
<a href="https://arxiv.org/abs/2404.12339">
<img src="./web_elements/paper.png" height="75pt"/>
</a>
</div>
<div class="col" style="width: 33%; height: 60pt;">
<a href="https://github.com/umautobots/SPOT">
<img src="./web_elements/github.png" height="75pt"/>
</a>
</div>
<div class="col" style="width: 33%; height: 60pt;">
<a href="https://umautobots.github.io/nsavp">
<img src="./web_elements/database.png" height="75pt"/>
</a>
</div>
</div>
<!-- Link name -->
<div class="row" style="margin-left:auto; margin-right:auto; max-width:800px; line-height: 10pt;">
<div class="col" style="width: 33%; font-size:12pt; font-weight: 600;">
<p>Preprint (will be updated with appendix soon)</p>
</div>
<div class="col" style="width: 33%; font-size:12pt; font-weight: 600;">
<p>Code</p>
</div>
<div class="col" style="width: 33%; font-size:12pt; font-weight: 600;">
<p>NSAVP Dataset</p>
</div>
</div>
</div>
<!-- Video -->
<!-- TODO: link to youtube video -->
<!-- Overview -->
<div class="container">
<h2 class="section_title">Abstract</h2>
<p class="body">Recognizing places from an opposing viewpoint during a return trip is a common experience for human drivers. However, the analogous robotics capability, visual place recognition (VPR) with limited field of view cameras under 180 degree rotations, has proven to be challenging to achieve. To address this problem, this paper presents Same Place Opposing Trajectory (SPOT), a technique for opposing viewpoint VPR that relies exclusively on structure estimated through stereo visual odometry (VO). The method extends recent advances in lidar descriptors and utilizes a novel double (similar and opposing) distance matrix sequence matching method. We evaluate SPOT on a publicly available dataset with 6.7-7.6 km routes driven in similar and opposing directions under various lighting conditions. The proposed algorithm demonstrates remarkable improvement over the state-of-the-art, achieving up to 91.7% recall at 100% precision in opposing viewpoint cases, while requiring less storage than all baselines tested and running faster than all but one. Moreover, the proposed method assumes no <i>a priori</i> knowledge of whether the viewpoint is similar or opposing, and also demonstrates competitive performance in similar viewpoint cases.
</p>
</div>
<!-- Citation -->
<div class="container">
<h2 class="section_title">Citation</h2>
<ul class="body" style="font-size: 10pt; line-height: 15pt;">
<ul style="padding-left:0"><div style="background-color:#e6e6e6;">
<p class="body" style="padding-left:0pt; font-size: 10pt;padding-bottom:0;text-indent:-3rem;margin-left:4rem;font-family:monospace;line-height: 1.2rem;">
@misc{carmichael2024spot, <br>
title={SPOT: Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints}, <br>
author={Spencer Carmichael and Rahul Agrawal and Ram Vasudevan and Katherine A. Skinner}, <br>
year={2024}, <br>
eprint={2404.12339}, <br>
archivePrefix={arXiv}, <br>
primaryClass={cs.RO}}
</div></li>
</ul>
</div>
</body>
</html>