-
Notifications
You must be signed in to change notification settings - Fork 0
/
tasks.html
194 lines (163 loc) · 6.21 KB
/
tasks.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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<title>DOTA2</title>
<!-- <link rel="stylesheet" href="http://cdn.static.runoob.com/libs/bootstrap/3.3.7/css/bootstrap.min.css"> -->
<link rel="stylesheet" href="bootstrap-3.3.7-dist/css/bootstrap.min.css">
<link rel="stylesheet" type="text/css" href="css/mystyle.css"
<script src="http://cdn.static.runoob.com/libs/jquery/2.1.1/jquery.min.js"></script>
<!-- <script src="http://cdn.static.runoob.com/libs/bootstrap/3.3.7/js/bootstrap.min.js"></script> -->
<style>
body {
background-color : rgb(250,250,250); /*UCLA bgcolor*/
}
</style>
</head>
<body>
<div class="container">
<div class="row">
<div class="span12">
<div class="content">
<h1 style="text-align:center; margin-top:60px; font-weight: bold">
Object Detection in Aerial Images: A Large-scale Benchmark and Challenges
</h1>
<p style="text-align:center; margin-bottom:15px; margin-top:20px; font-size: 18px">
<a href="http://captain.whu.edu.cn/dingjian.html">Jian Ding</a>,
<a href="https://cherubicxn.github.io/">Nan Xue</a>
<a href="http://captain.whu.edu.cn/xia_En.html">Gui-Song Xia</a>,
<a href="https://sites.google.com/site/michaelyingyang/">Micheal Ying Yang</a>
<a href="http://mclab.eic.hust.edu.cn/~xbai/">Xiang Bai</a>,
<a href="http://vision.cornell.edu/se3/people/serge-belongie/">Serge Belongie</a>,
<br>
<a href="http://www.cs.rochester.edu/u/jluo/">Jiebo Luo</a>,
<a href="http://www.dlr.de/caf/en/desktopdefault.aspx/tabid-5242/8788_read-933/sortby-lastname/">Mihai Datcu</a>,
<a href="http://www.dsi.unive.it/~pelillo/">Marcello Pelillo</a>,
<a href="http://www.lmars.whu.edu.cn/prof_web/zhangliangpei/rs/xueshu.htm">Liangpei Zhang</a>.
</p>
</div>
<div class="row">
<div class="span4">
</div>
<div class="span4">
<ul class="nav nav-tabs">
<!-- <li class="active"> -->
<li >
<a href="index.html">Home</a>
</li>
<li>
<a href="dataset.html">Dataset</a>
</li>
<li class="active">
<!-- <li> -->
<a href="tasks.html">Tasks</a>
</li>
<li>
<a href="evaluation.html">Evaluation</a>
</li>
<li>
<a href="results.html">Results</a>
</li>
<br />
</ul>
</div>
<div class="span4">
</div>
</div>
<h2>
Overview
</h2>
<p>
We propose two detection tasks. Task1 uses the initial annotation as ground truth.
Task2 uses the generated axis-aligned bounding boxes as ground truth.
The results from task2 are of great practical value.
We recommond you to test your algorithms by way of Task1.
</p>
<h2>Task1 - Detection with oriented bounding boxes <strong>(Recommended)</strong> </h2>
<p>
The aim of this task is to locate the ground object instances with an oriented bounding box. The
oriented bounding box follows the same format with the original annotation
</p>
<!-- <h3>
Evaluation Server
</h3>
<p>
For evaluation, you must registrate and submit on the
<a href="http://www.icdar2017chinese.site:5080/evaluation1/">Evaluation Server</a>
</p> -->
<h3>
Submission Format
</h3>
<p>
You will be asked to submit a zip file containing results for all test images to evaluate your results. The results are stored
in 18 files, <strong style="color:blue">"Task1_plane.txt, Task1_storage-tank.txt, ..."</strong>, each file contains all the results for a specific category.
Each file is in the following format:
</p>
<pre>
<code style="font-size:16px">
imgname score x<sub>1</sub> y<sub>1</sub> x<sub>2</sub> y<sub>2</sub> x<sub>3</sub> y<sub>3</sub> x<sub>4</sub> y<sub>4</sub>
imgname score x<sub>1</sub> y<sub>1</sub> x<sub>2</sub> y<sub>2</sub> x<sub>3</sub> y<sub>3</sub> x<sub>4</sub> y<sub>4</sub>
...
</code>
</pre>
<a href="submissionformat/example_task1.zip">An example submission of task1</a>
<h3>
Evaluation Protocol
</h3>
<p>
The evaluation protocol for oriented bounding box is a little different from the protocol in the
original PASCAL VOC. We use the intersection over the union area of two polygons(ground truth
and prediction) to calculate the IoU. The rest follows the PASCAL VOC.
</p>
<h2>
Task2 - Detection with horizontal bounding boxes
</h2>
<p>
Detecting object with horizontal bounding boxes is usual in many previous contests for object
detection. The aim of this task is to accurately localize the instance in terms of horizontal bounding
box with (xmin, ymin, xmax, ymax) format. In the task, the ground truths for training and testing are generated
by calculating the axis-aligned bounding boxes over original annotated bounding boxes.
</p>
<p>
</p>
<!-- <h3>
Evaluation Server
</h3>
<p>
For evaluation, you must registrate and submit on the
<a href="http://www.icdar2017chinese.site:5080/evaluation1/">Evaluation Server</a>
</p> -->
<h3>
Submission Format
</h3>
<p>
You will be asked to submit a zip file containing results for all test images to evaluate your results. The results are stored
in 18 files, <strong style="color:blue">"Task2_plane.txt, Task2_storage-tank.txt, ..."</strong>, each file contains all the results for a specific category.
</p>
<pre>
<code style="font-size:16px">
imgname score xmin ymin xmax ymax
imgname score xmin ymin xmax ymax
...
</code>
</pre>
<a href="submissionformat/example_task2.zip">An example submission of task2</a>
<h3>
Evaluation Protocol
</h3>
<p>
The evaluation protocol for horizontal bounding boxes follows the PASCAL VOC benchmark,
which uses mean Average Precision(<strong>mAP</strong>) as the primary metric.
</p>
</div>
</div>
<br>
<p align="center">
<a href="https://clustrmaps.com/site/1azfj" title="Visit tracker">
<img src="//www.clustrmaps.com/map_v2.png?d=nj-0YcGTgR29XUNgouMq7DH65Mt4ec4BvufmGUydZjg&cl=ffffff" />
</a>
</p>
<br>
</div>
</body>
</html>