-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
489 lines (468 loc) · 20.5 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
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
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
<!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" xml:lang="en" lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, maximum-scale=1.0, user-scalable=0" />
<link rel="stylesheet" type="text/css" href="webpage/bootstrap.min.css"/>
<script language="javascript" src="webpage/jquery.min.js"></script>
<script language="javascript" src="webpage/bootstrap.min.js"></script>
<link rel="stylesheet" type="text/css" href="webpage/cssReset.css"/>
<title>Welcome to Xiaoyu Zhu's homepage</title>
<meta name="description" content="Personal website for Xiaoyu Zhu at Carnegie Mellon University. Xiaoyu Zhu is a Ph.D. student in the Language Technologies Institute at Carnegie Mellon University.">
<meta name="keywords" content="Xiaoyu Zhu,CMU,computer vision">
<!-- Global site tag (gtag.js) - Google Analytics -->
<!-- <script async src="https://www.googletagmanager.com/gtag/js?id=UA-156016426-1"></script> -->
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-156016426-1');
</script>
</head>
<body>
<style type='text/css'>
body{
font-family:arial,"Microsoft YaHei",微软雅黑,宋体,Helvetica;
font-size:15px;
}
/*
div.content:
provide the content div in the middle
*/
body div.content{
/*width:1280px;*/
width:1000px;
margin:0 auto;
line-height:30px;
}
/*
header wrapper
*/
body div.header{
background-color:#2196F3;
}
body div.header > div.content{
padding:10px;
}
/*
footer1
*/
body div.footer1{
margin-top:30px;
background-color:#64B5F6;
}
/*
footer2
*/
body div.footer2{
background-color:#2196F3;
}
body div.footer > div.content{
padding:10px;
}
body div.footer1 > div.content{
padding:20px;
line-height:40px;
font-size:1.2em;
}
/*
utils css
*/
div.white-text{
color:white;
}
div.content > div.title{
padding:20px 0;
border-top:1px silver solid;
margin-top:30px;
font-size:2em;
font-weight:bold;
}
body a{
text-decoration:none;
}
div.content ul{
list-style: disc inside none;
}
div.content ol{
list-style: none inside none;
}
div.content li{
line-height:30px;
}
div.content div.float-right{
float:right;
}
</style>
<!-- css for bio -->
<style type="text/css">
/*
bio
*/
div.bio{
font-size:1.2em;
}
div.bio > div.left{
float:left;
width:300px;
}
div.bio > div.left > img.me{
max-height:190px;
margin:10px;
max-width:190px;
margin-top:40px;
margin-left:40px;
}
div.bio > div.right{
margin:0 0 0 310px;
min-height:300px;
}
div.bio > div.right > div.line.name{
padding:15px 0;
line-height:40px;
}
div.bio > div.right > div.name > span.name,div.bio > div.right > div.name > span.chineseName{
font-size:2em;
font-weight:bold;
}
div.bio > div.right > div.name > span.chinesesName{
font-family:"Microsoft YaHei",微软雅黑,宋体,Helvetica,arial;
}
div.bio > div.right > div.name > span.misc{
font-size:1.5em;
font-weight:bold;
}
div.bio > div.right > div.line.school{
padding:5px 0;
}
div.bio > div.right > div.line.office{
padding:5px 0;
}
</style>
<!-- quick link and intro -->
<style type="text/css">
div.quickLink{
min-height:70px;
}
div.quickLink > .block{
display:block;
float:left;
padding:10px 0px;
text-align:center;
border:1px silver solid;
border-radius:5px;
box-shadow:2px 2px 1px silver;
width:140px;
margin-right:50px;
margin-top:0px;
cursor:pointer;
}
</style>
<!-- research and education -->
<style type="text/css">
div.research > ul > li > span.title,div.research > ul > li > div.time,
div.education > ul > li > span.title,div.education > ul > li > div.time,
div.pro > ul > li > span.title{
font-weight:bold;
font-size:1.1em;
}
div.research > ul > li > div.info,
div.education > ul > li > div.info,
div.pro > ul > li > div.info{
padding-left:20px;
word-wrap:break-word;
}
</style>
<!-- publications -->
<style type="text/css">
div.publications > ol > li{
padding-bottom: 0px;
}
div.publications > ol > li > span.title{
font-weight:bold;
font-size:1.2em;
}
div.publications > ol > li > div.info{
padding-left:20px;
word-wrap:break-word;
}
div.publications > ol > li > div.info.italic{
font-style: italic;
}
div.publications div.imgblock{
float:left;
height:180px;
width:300px;
padding:0px;
margin-right:30px;
text-align: center;
}
div.publications div.imgblock > img{
max-width:100%;
max-height:100%;
}
img.press{
height:20px;
}
</style>
<div class="header">
<div class="content white-text">
Xiaoyu (Lesley) Zhu ([email protected])
</div>
</div>
<div class="content bio">
<!-- bio -->
<div class="left">
<img class='me' src="img/xylinkedin.png"></img>
</div>
<div class="right">
<div class="line name">
<span class="name">Xiaoyu Zhu</span>
<!-- <span class="misc">(M.S. Artificial Intelligence and Innovation)</span> -->
<!-- <span class="chineseName">梁俊卫</span> -->
</div>
<div class="line school">
Ph.D. Student in Artificial Intelligence
</div>
<div class="line school">
Language Technologies Institute
</div>
<div class="line school">
School of Computer Science
</div>
<div class="line school">
Carnegie Mellon University
</div>
</div>
</div>
<div class="content quickLink">
<a class="block" href="#publications">Publications</a>
<a class="block" href="https://scholar.google.com/citations?user=T4Dc5zEAAAAJ&hl=en">Google Scholar</a>
<a class="block" href="https://github.com/zgzxy001">Github</a>
<a class="block" href="https://www.linkedin.com/in/xiaoyuzhu3">Linkedin</a>
<a class="block" href="https://drive.google.com/file/d/1U6vtuQ16NlDmPRuwqXR8Cs9MYMfLxUNt/view?usp=sharing">Resume</a>
<!-- <a class="block" href="https://github.com/JunweiLiang" target="_blank">Code</a>
<a class="block" href="#teaching">Talks</a> -->
</div>
<div class="content intro">
<div class='title'>Introduction</div>
Greetings! I'm a Ph.D. student at Language Technologies Institute under the supervision of <a href="https://scholar.google.com/citations?user=Py54GcEAAAAJ&hl=en">Prof. Alexander Hauptmann</a>.<br/>
My research focuses on robust feature representation learning for visual perception and generation. I have been developing novel representation learning methods including: <span style="font-weight:bold"> (1) Contrastive/Siamese Learning </span> for video and 3D scene understanding; <span style="font-weight:bold"> (2) Masked Visual Modeling </span> for human action analysis from videos and 3D point clouds/meshes; and <span style="font-weight:bold">(3) Generatively Pretrained Vision-Language Models</span> for open-vocabulary 3D scene understanding and generation. I have published first-authored papers in top computer vision conferences such as CVPR, ICCV, and ECCV. My research has been featured by CMU News and Science Daily, and deployed by the Federal Emergency Management Agency. <br/>
<span style="font-weight:bold"> I'm actively seeking full-time industrial positions starting in Fall 2024 or Spring 2025. Please reach out via email if you are interested! </span>
</div>
<div class="content news">
<a name="news"></a>
<div class="title">News</div>
<ul>
<li>
[07/2024] One paper has been accepted by <span style="font-weight:bold">ECCV 2024</span>.
</li>
<li>
[02/2024] One paper has been accepted by <span style="font-weight:bold">SPIE DCS 2024 (Oral)</span>.
</li>
<li>
[02/2023] One paper has been accepted by <span style="font-weight:bold">CVPR 2023</span>.
</li>
<li>
[01/2023] One paper has been accepted by <span style="font-weight:bold">SPIE DCS 2023</span>.
</li>
<li>
[06/2022] Featured in a <a href="https://www.washingtonpost.com/investigations/interactive/2022/shireen-abu-akleh-death/?itid=lk_inline_manual_4"> news report</a> by Washington Post using shooter localization technologies.
<a href=""https://www.washingtonpost.com/investigations/interactive/2022/shireen-abu-akleh-death/?itid=lk_inline_manual_4"
<img class="press" src="resources/wapo.png"></img>
</a>
</li>
<li>
[07/2021] <a href="https://openaccess.thecvf.com/content/ICCV2021/papers/Zhu_Weakly_Supervised_3D_Semantic_Segmentation_Using_Cross-Image_Consensus_and_Inter-Voxel_ICCV_2021_paper.pdf"><span style="font-style: italic;">CIVANet</span></a> paper has been accepted by <span style="font-weight:bold">ICCV 2021</span>.
</li>
<li>
[03/2021] We developed an AI solution to improve cleaning schedules at Pittsburgh International Airport.
Press Coverage:
<a href="https://www.cs.cmu.edu/news/students-use-ai-improve-cleaning-schedules-pittsburgh-airport">
<img class="press" src="resources/cmu.png"></img>
</a>,
<a href="https://pittsburgh.cbslocal.com/2021/03/09/cmu-artificial-intelligence-cleaning-restrooms-pittsburgh-international-airport/">
<img class="press" src="resources/cbs.png"></img>
</a>,
<a href="https://blueskypit.com/2021/03/08/carnegie-mellon-university-teams-up-with-pit-for-smarter-restroom-cleaning/">
<img class="press" src="resources/pit.png"></img>
</a>
</li>
<li>
[09/2020] We won the <a href="https://www.herox.com/ASAPS1/community"><span style="font-style: italic;">Automated Streams Analysis for Public Safety Challenge</span></a> with a <a href="https://www.herox.com/ASAPS1/update/3483"><span style="font-style: italic;">$30k prize</span></a>.
</li>
<li>
[08/2020] Our automatic damage assessment system has been successfully tested on <a href="https://twitter.com/SCSatCMU/status/1299359834798616582?s=19"><span style="font-style: italic;">Hurricane Laura</span></a>.
Press Coverage:
<a href="https://www.cmu.edu/news/stories/archives/2020/august/drones-hurricane-damage.html">
<img class="press" src="resources/cmu.png"></img>
</a>,
<a href="https://pittsburgh.cbslocal.com/2020/08/29/cmu-amateur-drone-videos-social-media-storm-damage">
<img class="press" src="resources/cbs.png"></img>
</a>,
<a href="https://www.msn.com/en-us/news/us/cmu-amateur-drone-videos-posted-to-social-media-could-be-used-to-assess-storm-damage/ar-BB18vkAb">
<img class="press" src="resources/msn.png"></img>
</a>,
<a href="https://www.sciencedaily.com/releases/2020/08/200828081023.htm">
<img class="press" src="resources/science_daily.png"></img>
</a>,
<a href="https://www.azorobotics.com/News.aspx?newsID=11581">
<img class="press" src="resources/azo.png"></img>
</a>
</li>
<li>
[07/2020] <a href="https://arxiv.org/pdf/2006.16479.pdf"><span style="font-style: italic;">MSNet</span></a> paper has been accepted by <span style="font-weight:bold">WACV 2021</span> (with strong-accept).
</li>
</ul>
</div>
<div class="content publications">
<a name="publications"></a>
<div class="title">Publications</div>
<ol>
<li>
<div class="imgblock" style="height:250px;"><img src="img/diff2scene.png"></img></div>
<span class="title">Open Vocabulary 3D Semantic Segmentation with Text-to-Image Diffusion Models
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Zhou, H., Xing, P., Zhao L., Xu H., Liang J., Hauptmann, A., Liu, T., Gallagher, A.</div>
<div class="info"><span style="font-weight:bold">ECCV 2024</span></div>
<div class="stuff">
<a class="" href="https://arxiv.org/pdf/2407.13642" target="_blank">[Paper]</a>
<a class="" href="https://diff2scene.github.io/" target="_blank">[Project Page]</a>
</div>
<div style="clear:left"></div>
</li>
<li>
<div class="imgblock" style="height:250px;"><img src="img/cvpr_23.png"></img></div>
<span class="title">STMT: A Spatial-Temporal Mesh Transformer for MoCap-Based Action Recognition
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Huang, P.*, Liang, J.*, De Melo, C., Hauptmann, A.</div>
<div class="info"><span style="font-weight:bold">CVPR 2023</span></div>
<div class="stuff">
<a class="" href="https://openaccess.thecvf.com/content/CVPR2023/papers/Zhu_STMT_A_Spatial-Temporal_Mesh_Transformer_for_MoCap-Based_Action_Recognition_CVPR_2023_paper.pdf" target="_blank">[Paper]</a>
<a class="" href="https://github.com/zgzxy001/STMT" target="_blank">[Dataset/Code]</a>
</div>
<div style="clear:left"></div>
</li>
<li>
<div class="imgblock" style="height:250px;"><img src="img/spie_new.png"></img></div>
<span class="title">Leveraging Body Pose Estimation for Gesture Recognition Using Synthetic Data
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, De Melo, C., Hauptmann, A.</div>
<div class="info"><span style="font-weight:bold">SPIE DCS 2023</span></div>
<div class="stuff">
<a class="" href="https://www.spiedigitallibrary.org/conference-proceedings-of-spie/12529/125290Z/Leveraging-body-pose-estimation-for-gesture-recognition-in-human-robot/10.1117/12.2664030.pdf" target="_blank">[Paper]</a>
<a class="" href="https://github.com/zgzxy001/STMT" target="_blank">[Dataset/Code]</a>
</div>
<!-- <div class="keywords">
We introduce CIVA-Net, the first 3D volumetric segmentation model based on image-level supervision. It contains an inner-image consensus learning module to tackle the challenges brought by imaging limits, and an inter-voxel affinity learning module to predict segmentation with accurate boundaries of complex 3D structures.
</div> -->
<!-- <div class="stuff">
<a class="" href="https://arxiv.org/abs/1902.03748" target="_blank">[Paper]</a>
<a class="" href="camera_ready/future19.bib" target="_blank">[BibTex]</a>
<a class="" href="https://www.youtube.com/watch?v=NyrGxGoS01U" target="_blank">[Demo Video]</a>
<a class="" href="https://next.cs.cmu.edu" target="_blank">[Project Page/Code/Model]</a>
</div> -->
<div style="clear:both"></div>
</li>
<li>
<div class="imgblock" style="height:250px;"><img src="img/iccv_21.png"></img></div>
<span class="title">Weakly Supervised 3D Semantic Segmentation Using Cross-Image Consensus and Inter-Voxel Affinity Relations
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Chen, J., Zeng, X., Liang, J., Li, C., Liu, S., Xu, M.</div>
<div class="info"><span style="font-weight:bold">ICCV 2021</span></div>
<div class="stuff">
<a class="" href="https://openaccess.thecvf.com/content/ICCV2021/papers/Zhu_Weakly_Supervised_3D_Semantic_Segmentation_Using_Cross-Image_Consensus_and_Inter-Voxel_ICCV_2021_paper.pdf" target="_blank">[Paper]</a>
</div>
<!-- <div class="keywords">
We introduce CIVA-Net, the first 3D volumetric segmentation model based on image-level supervision. It contains an inner-image consensus learning module to tackle the challenges brought by imaging limits, and an inter-voxel affinity learning module to predict segmentation with accurate boundaries of complex 3D structures.
</div> -->
<!-- <div class="stuff">
<a class="" href="https://arxiv.org/abs/1902.03748" target="_blank">[Paper]</a>
<a class="" href="camera_ready/future19.bib" target="_blank">[BibTex]</a>
<a class="" href="https://www.youtube.com/watch?v=NyrGxGoS01U" target="_blank">[Demo Video]</a>
<a class="" href="https://next.cs.cmu.edu" target="_blank">[Project Page/Code/Model]</a>
</div> -->
<div style="clear:both"></div>
</li>
<li>
<div class="imgblock" style="height:250px;"><img src="img/msnet_new.png"></img></div>
<span class="title">MSNet: A Multilevel Instance Segmentation Network for Natural Disaster Damage Assessment in Aerial Videos
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Liang, J., Hauptmann, A.</div>
<div class="info"><span style="font-weight:bold">WACV 2021</span></div>
<div class="stuff">
<a class="" href="https://arxiv.org/pdf/2006.16479.pdf" target="_blank">[Paper]</a>
<a class="" href="https://github.com/zgzxy001/MSNET" target="_blank">[Dataset/Code]</a>
</div>
<!-- <div class="keywords">
We present the first natural disaster building damage assessment dataset. It provides the first quantitative benchmark for assessing damage assessment in aerial videos. We introduce MSNet, which contains Hierarchical Region Proposal Network and Score Refinement Network to tackle the challgense brought by huge view-point variances.
</div> -->
<div style="clear:both"></div>
</li>
<!-- <li>
<div class="imgblock" style="height:200px;"><img src="img/civa.png"></img></div>
<span class="title">Weakly Supervised 3D Semantic Segmentation Based on Cross-Image Consensus and Inter-Voxel Affinity Relations
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Chen, J., Zeng, X., Liang, J., Li, C., Zhang, J., Liu, S., Ban, X., Xu, M.</div>
<div class="info">CVPR 2021 (under review)</div>
<div class="keywords">
We introduce CIVA-Net, the first 3D volumetric segmentation model based on image-level supervision.
</div>
<div style="clear:both"></div>
</li>
<li>
<div class="imgblock" style="height:170px;" ><img src="img/unsupervise.png"></img></div>
<span class="title">Unsupervised 3D Semantic Segmentation Based on Multi-Level Self-Supervision
</span>
<div class="info text-success italic"><span style="font-weight:bold">Zhu, X.</span>, Zeng, X., Li, C., Liu, S., Ban, X., Xu, M.</div>
<div class="info">Nature Machine Intelligence (in preparation)<br/></div>
<div class="keywords">
We present an unsupervised segmentation method based on multi-level self-supervision. The feature maps produced by each level of self-supervision are ensembled at the end to produce robust and consistent semantic segmentation.
</div>
<div style="clear:both"></div>
</li> -->
</ol>
</div>
<!-- <div class="content honors" name='honors'>
<a name="honors"></a>
<div class="title">Awards</div>
<ul>
<li> Winner, Automated Streams Analysis for Public Safety Challenge - $30k prize, 2020</li>
<li> Outstanding Alumni of BJTU&RIT, 2020</li>
<li> Cohort Commencement Speaker, 2019</li>
<li> Highest Honor of RIT, 2019</li>
<li> National Scholarship of China, 2017</li>
<li> Global Scholar, 2017</li>
<li> First Class Scholarship/Merit Student/Dean's List of BJTU&RIT, 2015-2019</li>
</ul>
</div>-->
<div class="content pro">
<a name="Media"></a>
<div class="title">Selected Media</div>
<ul>
<li> <strong>Carnegie Mellon University News.</strong> Students Use AI To Improve Cleaning Schedules at Pittsburgh Airport, March 11, 2021. </li>
<li> <strong>Carnegie Mellon University News.</strong> Amateur Drone Videos Could Aid in Natural Disaster Damage Assessment, August 28, 2020. </li>
<li> <strong>Science Daily.</strong> Amateur Drone Videos Could Aid in Natural Disaster Damage Assessment, August 28, 2020. </li>
<li> <strong>CBS.</strong> CMU: Amateur Drone Videos Posted To Social Media Could Be Used To Assess Storm Damage. August 29, 2020.</li>
<li> <strong>Microsoft News.</strong> CMU: Amateur Drone Videos Posted To Social Media Could Be Used To Assess Storm Damage. August 31, 2020.</li>
<li> <strong>Yahoo News.</strong> CMU Developing Program To Assess Hurricane Damage. August 29, 2020.</li>
<li> <strong>AZO Robotics.</strong> New AI System Helps Detect Damage Caused to Buildings by Hurricanes. August 31, 2020.</li>
</ul>
</div>
<div class="footer footer1">
<div class="content white-text">
Informedia Lab, Language Technologies Institute <br/>
School of Computer Science <br/>
Carnegie Mellon University
</div>
<div class="footer footer2">
<div class="content white-text">
Designed by <a style="color:white" href="https://www.cs.cmu.edu/~junweil/">Junwei Liang</a> in CMU.
</div>
</div>
<!--
a Junwei Liang's production
contact: [email protected]
-->
</body>
</html>