forked from x-decoder-vl/x-decoder-vl.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
455 lines (411 loc) · 19.4 KB
/
index.html
File metadata and controls
455 lines (411 loc) · 19.4 KB
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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="Deformable Neural Radiance Fields creates free-viewpoint portraits (nerfies) from casually captured videos.">
<meta name="keywords" content="Nerfies, D-NeRF, NeRF">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>X-Decoder: Generalized Decoding for Pixel, Image and Language</title>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-PYVRSFMDRL"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'G-PYVRSFMDRL');
</script>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="icon" href="./static/images/icon.png">
<link rel="stylesheet" href="./static/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">X-Decoder:</h1>
<h2 class="title is-2 publication-title">Generalized Decoding for Pixel, Image and Language</h2>
<div class="is-size-5">
<span class="author-block">
<a href="https://maureenzou.github.io/" style="color:#f68946;font-weight:normal;">Xueyan Zou<sup>*</sup>
</a>,
</span>
<span class="author-block">
<a href="https://zdou0830.github.io/" style="color:#F2A900;font-weight:normal;">Zi-Yi Dou<sup>*</sup></a>,</span>
<span class="author-block">
<a href="https://jwyang.github.io/" style="color:#00A4EF;font-weight:normal;">Jianwei Yang<sup>*⚑</sup></a>,
</span>
<span class="author-block">
<a href="https://zhegan27.github.io/" style="color:#008AD7;font-weight:normal;">Zhe Gan</a>,
</span>
<span class="author-block">
<a href="https://www.microsoft.com/en-us/research/people/linjli/" style="color:#008AD7;font-weight:normal;">Linjie Li</a>,
</span>
<span class="author-block">
<a href="https://chunyuan.li/" style="color:#00A4EF;font-weight:normal;">Chunyuan Li</a>,
</span>
<span class="author-block">
<a href="https://sites.google.com/site/xiyangdai/" style="color:#008AD7;font-weight:normal;">Xiyang Dai</a>,
</span>
<span class="author-block">
<a href="https://harkiratbehl.github.io/" style="color:#00A4EF;font-weight:normal;">Harkirat Behl</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=vJWEw_8AAAAJ&hl=en" style="color:#008AD7;font-weight:normal;">Jianfeng Wang</a>,
</span>
<span class="author-block">
<a href="https://www.microsoft.com/en-us/research/people/luyuan/" style="color:#008AD7;font-weight:normal;">Lu Yuan</a>,
</span>
<span class="author-block">
<a href="https://vnpeng.net/" style="color:#F2A900;font-weight:normal;">Nanyun Peng</a>,
</span>
<span class="author-block">
<a href="https://www.microsoft.com/en-us/research/people/lijuanw/" style="color:#008AD7;font-weight:normal;">Lijuan Wang</a>,
</span>
<span class="author-block">
<a href="https://pages.cs.wisc.edu/~yongjaelee/" style="color:#f68946;font-weight:normal;">Yong Jae Lee<sup>☨</sup></a>,
</span>
<span class="author-block">
<a href="https://www.microsoft.com/en-us/research/people/jfgao/" style="color:#00A4EF;font-weight:normal;">Jianfeng Gao<sup>☨</sup></a>
</span>
</div>
<br>
<div class="is-size-5 publication-authors">
<span class="author-block"><b style="color:#f68946; font-weight:normal">▶ </b> University of Wisconsin-Madison; </b></span>
<span class="author-block"><b style="color:#F2A900; font-weight:normal">▶ </b>UCLA; </span>
<span class="author-block"><b style="color:#00A4EF; font-weight:normal">▶ </b>Microsoft Research, Redmond; </span>
<span class="author-block"><b style="color:#008AD7; font-weight:normal">▶ </b>Microsoft Cloud & AI </span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>*</sup>Equal Technical Contribution, </span>
<span class="author-block"><sup>☨</sup>Equal Advisory Contribution, </span>
<span class="author-block"><sup>⚑</sup>Project Lead </span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a href="https://arxiv.org/pdf/2212.11270.pdf" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
<span class="link-block">
<a href="https://github.com/microsoft/X-Decoder" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
<span class="link-block">
<a href="https://huggingface.co/spaces/xdecoder/Demo"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="far fa-images"></i>
</span>
<span>Hugging Face Demo</span>
</a>
</span>
<span class="link-block">
<a href="https://youtu.be/wYp6vmyolqE"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-youtube"></i>
</span>
<span>Video</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<img id="teaser" width="120%" src="images/tittle_fig.gif">
<h2 class="subtitle has-text-centered">
<p style="font-family:Times New Roman"><b>Figure 1. X-Decoder is a single model trained to support a wide range of vision and vision-language tasks.</b></p>
</h2>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Abstract</h2>
<div class="content has-text-justified">
<p>
We present X-Decoder, a generalized decoding pipeline that can predict pixel-level segmentation and language tokens seamlessly.
X-Decoder takes as inputs two types of queries: (i) generic non-semantic queries and
(ii) semantic queries induced from text inputs, to decode different pixel-level and token-level outputs in the same semantic space. With such a novel design,
<b>X-Decoder is the first work that provides a unified way to support all types of image segmentation and a variety of vision-language (VL) tasks</b>.
After pretraining on a mixed set of a limited amount of segmentation data and million of image-text pairs, X-Decoder exhibits strong transferability to a wide range of downstream tasks in both zero-shot and finetuning settings.
Notably, it achieves <b>(1) state-of-the-art open-vocabulary segmentation and referring
segmentation on 10 settings of 7 datasets; (2) better or competitive finetuned performance to other generalist and specialist models on segmentation and VL tasks;
and (3) flexibility for efficient finetuning and novel task composition (e.g., referring captioning = referring segmentation + captioning).</b>
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
<br>
<br>
<!-- Paper video. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Video</h2>
<div class="publication-video">
<iframe src="https://user-images.githubusercontent.com/11957155/209045241-916ccf73-d29d-4637-8502-027d3420875c.mp4"
frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
</div>
</div>
</div>
<!--/ Paper video. -->
<br>
<br>
<!-- Paper Model. -->
<div class="columns is-centered has-text-centered">
<div class="column is-six-fifths">
<h2 class="title is-3">Model</h2>
<div class="content has-text-justified">
<p>
<b>Our X-Decoder is unique for three critical designs</b>:
</p>
<ul>
<li>It has two types of queries (latent queries and text queries) and outputs (semantic outputs and pixel-level outputs).</li>
<li>It uses a single text encoder for all text corpus, ranging from class concepts, referring phrases to image captions.</li>
<li>It decouples image and text encoder to accomadate cross-image tasks (e.g., image-text retrieval) and within-image tasks (e.g., segmentation and captioning).</li>
</ul>
</div>
<img id="model" width="50%" src="images/method1.gif">
<h3 class="subtitle has-text-centered">
<p style="font-family:Times New Roman"><b>Figure 2. X-Decoder overall architecture.</b></p>
</h3>
<br>
<br>
<div class="content has-text-justified">
<p>
<b>X-Decoder can be used to unify a variety of vision and vision-language tasks</b>:
</p>
<ul>
<li><b>Generic Segmentation</b>: Instance, semantic and panoptic segmentation, all supporting open-vocabulary and zero-shot.</li>
<li><b>Referring Segmentation</b>: Refer to specific image segment given arbitary textual queries from text encoder. </li>
<li><b>Image-Text Retrieval</b>: Decoupled image and text representation extraction and dot-product for computing the similarities. </li>
<li><b>Image Captioning</b>: Decode textual tokens using exactly the same decoder in autoregressive manner.</li>
</ul>
</div>
<img id="model" width="100%" src="images/method2.gif">
<h3 class="subtitle has-text-centered">
<p style="font-family:Times New Roman"><b>Figure 3. Unify different types of vision and vision-language tasks with a single X-Decoder.</b></p>
</h3>
</div>
</div>
</div>
<br>
<br>
<!--/ Paper video. -->
</div>
</section>
<section class="section">
<!-- Results. -->
<div class="columns is-centered has-text-centered">
<div class="column is-six-fifths">
<h2 class="title is-3">Results</h2>
</div>
</div>
</div>
<!--/ Results. -->
<div class="container is-max-desktop">
<!-- Generic Segmentation. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Zero-Shot Segmentation</h2>
<img id="teaser" width="120%" src="images/bar_chart.png">
<h1>
<p style="font-family:Times New Roman"><b>Chart 1. Left: Zero-shot Segmentation Result Compared with SoTA; Right: Finetuned Segmentation Result Compared with SoTA.</b>
</h1>
</div>
</div>
<div class="container is-max-desktop">
<div class="column is-full-width">
<h2 class="title is-4">Segmentation In the Wild</h2>
<img id="teaser" width="120%" src="images/zero-shot-seginw.png">
<h1>
<p style="font-family:Times New Roman"><b>Chart 2. Zero-shot instance segmentation performance on SeginW benchmark.</b>
</h1>
</div>
</div>
<section class="section">
<!-- Results. -->
<div class="columns is-centered has-text-centered">
<div class="column is-six-fifths">
<h2 class="title is-3">Demo</h2>
</div>
</div>
</div>
<!--/ Results. -->
<div class="container is-max-desktop">
<!-- Generic Segmentation. -->
<!-- <h1><b>Referring Image Editing and Inpainting</b></h1> -->
<h2 class="title is-4">Rerferring Image Editing and Inpainting</h2>
<div class="column is-full-width">
<div class="columns is-centered">
<img id="teaser" width="100%" src="images/inpaint.gif">
</div>
<div class="columns is-centered">
<h1>
<p style="font-family:Times New Roman"><b>Demo. Left: Green Apple -> Rabbit, Darkest Red Apple -> Red Pear, Table -> Plate. Middle: Dog Removal, Right: Sky -> Mountain</b>
</h1>
</div>
</div>
<!-- <div class="column is-full-width">
<h1><b>Referring Imaeg Inpainting</b></h1>
<img id="teaser" width="100%" src="images/remove-dog.gif">
<h1>
<p style="font-family:Times New Roman"><b>Demo 2. Remove the dog.</b>
</h1>
</div> -->
</div>
<section class="section">
<!-- Results. -->
<div class="columns is-centered has-text-centered">
<div class="column is-six-fifths">
<h2 class="title is-3">Visualizations</h2>
</div>
</div>
</div>
<!--/ Results. -->
<div class="container is-max-desktop">
<!-- Generic Segmentation. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Zero-Shot Segmentation</h2>
<!-- <h1><b>Zero-Shot Segmentation</b></h1> -->
<img id="teaser" width="120%" src="images/zero-shot-generic-image-segmentation.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 4. Zero-shot semantic segmentation with pretrained X-Decoder on 10 settings of 7 datasets.</b>
</h1>
</div>
</div>
<!--/ Generic Segmentation. -->
<!-- Video Segmentation. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Zero-shot Segmentation for Videos</h2>
<!-- <h1><b>Zero-shot Segmentation for Videos</b></h1> -->
<img id="teaser" width="120%" src="images/zs_video_seg.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 5. Zero-shot panoptic segmentation for <a href="https://youtube-vos.org/">YouTubeVOS</a> video dataset.</b>
</h1>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Referring Segmentation for Videos</h2>
<!-- <h1><b>Referring Segmentation for Videos</b></h1> -->
<img id="teaser" width="120%" src="images/referring-segmentation.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 6. Zero-shot referring segmentation for <a href="https://youtube-vos.org/">YouTubeVOS</a> video dataset.</b>
</h1>
</div>
</div>
<!-- VL Tasks. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Image Captioning</h2>
<!-- <h1><b>Image Captioning</b></h1> -->
<img id="teaser" width="120%" src="images/zero-shot-image-captioning-video-images.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 7. Zero-shot image captioning on YoutubeVOS video frames.</b>
</h1>
</div>
</div>
<!--/ VL Tasks. -->
<!-- Region Captioning. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Referring Image Captioning</h2>
<!-- <h1><b>Referring Image Captioning</b></h1> -->
<img id="teaser" width="120%" src="images/referring-captioning.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 8. Zero-shot referring image captioning on COCO 2017 val images (pink regions are referred).</b>
</h1>
</div>
</div>
<!--/ Region Captioning. -->
<!-- Region Editing. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Referring Image Editing</h2>
<!-- <h1><b>Referring Image Editing</b></h1> -->
<img id="teaser" width="120%" src="images/referring-image-editing.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 9. Zero-shot referring image editing by combining X-Decoder with Stable Diffusion inpainting.</b>
</h1>
</div>
</div>
<!--/ Region Editing. -->
<!-- Segmentation in the Wild. -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-4">Segmentation in the Wild</h2>
<img id="teaser" width="100%" src="images/seginw_examples.png">
<h1>
<p style="font-family:Times New Roman"><b>Figure 10. Ground Truth visualization of segmentation in the wild datasets from <a href="https://universe.roboflow.com/">Roboflow</a> for a wider evaluation. </b>
</h1>
</div>
</div>
<!--/ Segmentation in the Wild. -->
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>
@article{zou2022xdecoder,
author = {Zou, Xueyan and Dou, Zi-Yi and Yang, Jianwei and Gan, Zhe and Li, Linjie and Li, Chunyuan and Dai, Xiyang and Wang, Jianfeng and Yuan, Lu and Peng, Nanyun and Wang, Lijuan and Lee, Yong Jae and Gao, Jianfeng},
title = {Generalized Decoding for Pixel, Image and Language},
publisher = {arXiv:2212.11270v1},
year = {2022},
}
</code></pre>
</div>
</section>
<section class="section" id="Acknowledgement">
<div class="container is-max-desktop content">
<h2 class="title">Acknowledgement</h2>
<p>
This website is adapted from <a
href="https://github.com/nerfies/nerfies.github.io">Nerfies</a>, licensed under a <a rel="license"
href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</section>
</body>
</html>