-
Notifications
You must be signed in to change notification settings - Fork 0
/
tasks.html
342 lines (309 loc) · 16.6 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
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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<meta name="description" content="">
<meta name="author" content="">
<title>Learning to Understand Aerial Images (LUAI)</title>
<!-- Bootstrap core CSS -->
<link href="vendor/bootstrap/css/bootstrap.min.css" rel="stylesheet">
<!-- Custom fonts for this template -->
<link href="vendor/fontawesome-free/css/all.min.css" rel="stylesheet" type="text/css">
<link href='https://fonts.googleapis.com/css?family=Lora:400,700,400italic,700italic' rel='stylesheet' type='text/css'>
<link href='https://fonts.googleapis.com/css?family=Open+Sans:300italic,400italic,600italic,700italic,800italic,400,300,600,700,800' rel='stylesheet' type='text/css'>
<!-- Custom styles for this template -->
<link href="css/myclean-blog.min.css" rel="stylesheet">
<link href="css/common.css" rel="stylesheet">
</head>
<body>
<!-- Navigation -->
<nav class="navbar navbar-expand-lg navbar-light fixed-top" id="mainNav">
<div class="container mContainer">
<button class="navbar-toggler navbar-toggler-right" type="button" data-toggle="collapse" data-target="#navbarResponsive" aria-controls="navbarResponsive" aria-expanded="false" aria-label="Toggle navigation">
Menu
<i class="fas fa-bars"></i>
</button>
<div class="collapse navbar-collapse" id="navbarResponsive">
<ul class="navbar-nav ml-auto m-navbar">
<li class="nav-item">
<a class="nav-link" href="challenge.html" style="font-size:14px">Home</a>
</li>
<li class="nav-item">
<a class="nav-link" href="registration.html" style="font-size:14px">Registration</a>
</li>
<li class="nav-item">
<a class="nav-link" href="dataset.html" style="font-size:14px">Dataset</a>
</li>
<li class="nav-item">
<a class="nav-link" href="tasks.html" style="font-size:14px">Tasks</a>
</li>
<li class="nav-item">
<a class="nav-link" href="evaluation.html" style="font-size:14px">Evaluation</a>
</li>
<!--
<li class="nav-item dropdown">
<a class="nav-link dropdown-toggle" href="#" id="navbarDropdown" style="font-size:14px" role="button" data-toggle="dropdown"
aria-haspopup="true" aria-expanded="false">
Challenge
</a>
<div class="dropdown-menu" aria-labelledby="navbarDropdown">
<a class="dropdown-item" href="registration.html">Registration</a>
<a class="dropdown-item" href="dataset.html">Dataset</a>
<a class="dropdown-item" href="tasks.html">Tasks</a>
<a class="dropdown-item" href="evaluation.html">Evaluation</a>
<a class="dropdown-item" href="results.html">Results</a>
</div>
</li> -->
<li class="nav-item">
<a class="nav-link" href="results.html" style="font-size:14px"> Results </a>
</li>
<li class="nav-item">
<a class="nav-link" href="index.html" style="font-size:14px">Workshop</a>
</li>
</ul>
</div>
</div>
</nav>
<!-- Page Header -->
<!-- <header class="masthead m-height" style="background-image: url('images/bgGIF.gif');">
<div class="overlay" style="background-color: #212529;opacity: 0.85"></div>
<div class="container">
<div class="row">
<div class="col-lg-8 col-md-10 mx-auto">
<div class="site-heading m-head">
<h2 style="text-align:center; margin-top:60px; font-weight: bold;">
Challenge-2021 on
</h2>
<h1 style="text-align:center; font-weight: bold; font-size: 38px;color:#FF9900">
Learning to Understand Aerial Images
</h1>
<h2 style="text-align:center; font-weight: bold; font-style: italic">
<span class="subheading" style="text-align:center; font-weight:bold; font-style: italic"></span>
October 11, 2021, Montreal, Canada.
</h2>
</div>
</div>
</div>
</div>
</header> -->
<header class="masthead m-height" style="background-image: url('images/bgGIF.gif');">
<div class="overlay" style="background-color: #212529;opacity: 0.85"></div>
<div class="container">
<div class="row">
<div class="col-lg-8 col-md-10 mx-auto">
<div class="site-heading m-head">
<h2 style="text-align:center; margin-top:1.7em ;font-weight: bold;">
Challenge-2021 on
</h2>
<h1 style="text-align:center; font-weight: bold ; color:#FF9900">
<nobr>Learning to Understand Aerial Images</nobr>
</h1>
</h3 style="text-align:center; font-weight: bold; font-style: italic">
October 11, 2021, Montreal, Canada.
</h3>
</div>
</div>
</div>
</div>
</header>
<!-- Main Content -->
<div class="container">
<div class="row">
<div class="col-lg-10 col-md-10 mx-auto">
<h2>
Overview
</h2>
<p>
We propose three tasks. Task1 is to detect instances with oriented bounding boxes. Task2 is to detect instances with horizontal bounding boxes. Task3 is to detect semantic labels for each pixel. Task 1 and
Task 2 is based on DOTA-v2.0. Task 3 is based on GID-15.
You can use the provided train/val data to train and validate your detector. Validation data may also be used for training when submitting results on the test set. External data of any form is allowed. But must be reported during submission. Fine-tuning
models that are pretrained on ImageNet or COCO are also allowed.
</p>
<h2>
<b>Task1</b> - Detection with oriented bounding boxes
<!--<strong>(Recommended)</strong> -->
</h2>
<p>
The purpose of this task is to localize the ground object instances with an oriented bounding box. The oriented bounding box follows the same format with the original annotation {(x
<sub>i</sub>, y
<sub>i</sub>), i = 1,2,3,4}.
</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 <a href="example_Task1.zip">(example of task1)</a> 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>
<div class="alert alert-secondary" role="alert" style="font-size:18px;font-style: italic;font-family:'Times New Roman', Times, serif">
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> <br> 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> <br> ...
<!-- 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> category score
<br> 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> category score
<br> ... -->
</div>
<h3>
Evaluation Protocol
</h3>
<p>
The evaluation protocol for the 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>
<b>Task2</b> - 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 (x, y, w, h) 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>
<h3>
Submission Format
</h3>
<p>
You will be asked to submit a zip file <a href="example_Task2.zip">(example of task2)</a> containing results for all test images to evaluate your results. The results are stored in 16 files,
<strong style="color:blue">"Task2_plane.txt, Task2_storage-tank.txt, ..."</strong>, each file contains all the results for a specific category. The format of the results is:
</p>
<div class="alert alert-secondary" role="alert" style="font-size:18px;font-style: italic;font-family:'Times New Roman', Times, serif">
<!-- x y w h category score
<br> x y w h category score
<br> ... -->
imgname score xmin ymin xmax ymax <br> imgname score xmin ymin xmax ymax <br> ...
</div>
<!-- <a href="submissionformat/example_task2.rar">An example submission of task1</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>
<h2>
<b>Task3</b> - Semantic Segmentation
</h2>
<p>
The aim of this task is to give the semantic category for each pixel in aerial images.
</p>
<h3>
Submission Format
</h3>
<p>
Participants will be asked to submit a zip file <a href="http://47.108.71.49:8008/media/example.zip">(example of task3)</a> containing results (stored in ".png" format) for all test images.
Each ".png" file should have the same name as the corresponding tested image.
Image dimensions of ".png" files must be equal to input RGB image dimensions.
<!-- For example, expected result format for each image is a ".png" image with the same resolution as the tested image.-->
Each category is represented by a specific value:
</p>
<div class="alert alert-secondary" role="alert" style="font-size:18px;font-style: italic;font-family:'Times New Roman', Times, serif">
Paddyfield : 5<br> Urbanresidential: 2<br>
</div>
<h3>
Evaluation Protocol
</h3>
<p>
The evaluation protocol adopts the mIoU.
</p>
<!-- <h2>
<b>Task3</b> - Jointly object detection and orientation estimation for movable instances
</h2>
<p>
This task aims to estimate the orientation for movable instances(vehicles, planes, and ships), which is important when applied
to tracking. To make it clear, in this task, each instance's location and orientation is represented
by (x, y, w, h, θ) transferred from {(x
<sub>i</sub>, y
<sub>i</sub>), i = 1,2,3,4}.
</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 format of the results is:
</p>
<div class="alert alert-secondary" role="alert" style="font-size:18px;font-style: italic;font-family:'Times New Roman', Times, serif">
x y w h Θ category score
<br> x y w h Θ category score
<br> ...
</div>
<h3>
Evaluation Protocol
</h3>
<p>
Note that the evaluation protocol of this task is slightly different from that of Task 2. In Task 2, if the IoU between the
predicted box and ground truth is more than a certain threshold, it is assigned to be true positive(TP).
While in Task3, in addition to requiring IoU to be more than a threshold, the difference in angle is
required to be less than a certain threshold. As a result, the mAP for moveable classes in Task2 is upbound
of Task3. There is a similar metric in PASCAL3D+ which was called Average Viewpoint Precision(
<strong>AVP</strong>).
</p> -->
</div>
</div>
</div>
<hr>
<!-- Footer -->
<footer>
<div class="container">
<div class="row">
<div class="col-lg-8 col-md-10 mx-auto">
<ul class="list-inline text-center">
<li class="list-inline-item">
<a href=" https://twitter.com/intent/tweet?text=https%3A//captain-whu.github.io/LUAI2021">
<span class="fa-stack fa-lg">
<i class="fas fa-circle fa-stack-2x"></i>
<i class="fab fa-twitter fa-stack-1x fa-inverse"></i>
</span>
</a>
</li>
<li class="list-inline-item">
<!-- <a href="https://www.facebook.com/sharer/sharer.php?u=https%3A//captain-whu.github.io/DOAI2019"> -->
<a href="https://facebook.com/sharer.php?display=page&u=https%3A//captain-whu.github.io/LUAI2021">
<span class="fa-stack fa-lg">
<i class="fas fa-circle fa-stack-2x"></i>
<i class="fab fa-facebook-f fa-stack-1x fa-inverse"></i>
</span>
</a>
</li>
</ul>
<p class="copyright text-muted">Copyright ©
<a href="http://captain.whu.edu.cn/">CAPTAIN</a>
</p>
</div>
</div>
</div>
</footer>
<!-- Bootstrap core JavaScript -->
<script src="vendor/jquery/jquery.min.js"></script>
<script src="vendor/bootstrap/js/bootstrap.bundle.min.js"></script>
<!-- Custom scripts for this template -->
<script src="js/clean-blog.min.js"></script>
</body>
</html>