-
Notifications
You must be signed in to change notification settings - Fork 1
/
index.html
527 lines (483 loc) · 20.4 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
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description" content="EMMA: Your Text-to-Image Diffusion Model Can Secretly Accept Multi-Modal Prompts">
<meta name="keywords" content="EMMA">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title> EMMA: Your Text-to-Image Diffusion Model Can Secretly Accept Multi-Modal Prompts</title>
<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="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/icon.png">
<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>
<style>
.gif-row {
display: flex;
justify-content: center;
/* Center the GIFs horizontally */
align-items: center;
/* Center the GIFs vertically if they have different heights */
gap: 0px;
/* Space between GIFs */
}
.gif-row img {
max-width: 30%;
/* Ensure the GIFs are responsive */
}
.gif-column {
display: flex;
flex-direction: column;
align-items: center;
}
.gif-column img {
margin: 0px 0;
max-width: 30%;
/* Ensure the GIFs are responsive */
}
</style>
</head>
<!--
[**Yucheng Han**$^{\ast}$](http://tingxueronghua.github.io), [**Chi Zhang**$^{\ast\dag}$](https://icoz69.github.io/), [Xin Chen](https://chenxin.tech/), [Xu Yang](https://cse.seu.edu.cn/2021/1126/c23024a392593/page.htm), [Zhibin Wang](https://openreview.net/profile?id=~Billzb_Wang1)
<br>
[Gang Yu](https://www.skicyyu.org/), [Bin Fu](https://openreview.net/profile?id=~BIN_FU2), [Hanwang Zhang](https://personal.ntu.edu.sg/hanwangzhang/) -->
<body>
<!-- Yucheng Han, Chi Zhang, Xin Chen, Xu Yang, Billzb Wang, Gang Yu, BIN FU, Hanwang Zhang -->
<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">EMMA: Your Text-to-Image Diffusion Model Can Secretly Accept Multi-Modal Prompts
</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://tingxueronghua.github.io/">Yucheng Han*</a>,</span>
<span class="author-block">
<a href="https://wrong.wang/">Rui Wang*†</a>,</span>
<span class="author-block">
<a href="https://icoz69.github.io/">Chi Zhang*✦</a>,</span>
<span class="author-block">
<a href="https://tencentqqgylab.github.io/EMMA">Juntao Hu</a>,</span>
<span class="author-block">
<a href="https://tencentqqgylab.github.io/EMMA">Pei Cheng</a>,</span>
<span class="author-block">
<a href="https://tencentqqgylab.github.io/EMMA">Bin Fu</a>,</span>
<span class="author-block">
<a href="https://personal.ntu.edu.sg/hanwangzhang/">Hanwang Zhang</a></span>
<div class="is-size-5 publication-authors">
<span class="author-block">Nanyang Technological University, Tencent</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup></sup>(* Equal contributions, † Project Lead, ✦ Corresponding Author)</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2406.09162.pdf"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<span class="link-block">
<a href="https://arxiv.org/abs/2406.09162"
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>
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/TencentQQGYLab/ELLA"
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>
<!-- Huggingface Link. -->
<span class="link-block">
<a href="https://tencentqqgylab.github.io/EMMA"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Huggingface Demo (Coming soon)</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<!-- <figure>
<img src="./static/teaser.png" class="interpolation-image"
alt="Interpolate start reference image." />
<figcaption>
</figcaption>
</figure> -->
<div style="display: flex; justify-content: center; align-items: center;">
<video id="teaser" controls="" loop="" playsinline="">
<source src="./DreamBooth_files/video_teaser_480.mp4" type="video/mp4" codecs="avc1.42E01E, mp4a.40.2">
</video>
</div>
</div>
</div>
</div>
</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>
Recent advancements in image generation have enabled the creation of high-quality images from text conditions.
However, when facing multi-modal conditions, such as text combined with reference appearances, existing methods
struggle to balance multiple conditions effectively, typically showing a preference for one modality over others.
To address this challenge, we introduce EMMA, a novel image generation model accepting multi-modal prompts built
upon the state-of-the-art text-to-image (T2I) diffusion model, ELLA. EMMA seamlessly incorporates additional
modalities alongside text to guide image generation through an innovative Multi-modal Feature Connector design,
which effectively integrates textual and supplementary modal information using a special attention mechanism. By
freezing all parameters in the original T2I diffusion model and only adjusting some additional layers, we reveal
an interesting finding that the pre-trained T2I diffusion model can secretly accept multi-modal prompts. This
interesting property facilitates easy adaptation to different existing frameworks, making EMMA a flexible and
effective tool for producing personalized and context-aware images and even videos. Additionally, we introduce a
strategy to assemble learned EMMA modules to produce images conditioned on multiple modalities simultaneously,
eliminating the need for additional training with mixed multi-modal prompts. Extensive experiments demonstrate the
effectiveness of EMMA in maintaining high fidelity and detail in generated images, showcasing its potential as a
robust solution for advanced multi-modal conditional image generation tasks.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</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">Method</h2>
<div class="content has-text-justified">
<p>
Our proposed EMMA is built upon the state-of-the-art text-conditioned
diffusion model ELLA, which trains a transformer-like module, named Perceiver Resampler, to
connect text embeddings from pre-trained text encoders and pre-trained diffusion models for better text-guided
image generation.
ELLA has strong text-to-image generation ability, and our proposed EMMA could merge information from other
modalities into text features for
guidance.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/emma_pipeline.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</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"></h2>
<div class="content has-text-justified">
<p>
In detail, to control the image generation process by modalities beyond text, EMMA incorporates our proposed
Assemblable Gated Perceiver Resampler (AGPR), which leverages cross-attention to inject information from
additional modalities beyond texts. In our design, the AGPR blocks are strategically interleaved with the
blocks
of the Perceiver Resampler of ELLA. This arrangement ensures an effective integration of multi-modal
information.
During training, we freeze the raw modules of ELLA to maintain the control ability of text conditions.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/emma_methods.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</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"></h2>
<div class="content has-text-justified">
<p>
Notably, EMMA is inherently designed to handle multi-modal prompts as conditions, allowing for the
straightforward combination of different multi-modal configurations. This is achieved by the gate mechanism in
our
AGPR, which could control the way of injecting information from other modalities into the textual features.
This
advantage enables diverse and complex inputs to be synthesized into a unified generation framework without the
need for additional training.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/emma_ensemble_methods.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</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">Image Generation with text + portrait conditions</h2>
<div class="content has-text-justified">
<p>Here, we present additional images generated by EMMA under text + portrait conditions. Various portraits, each
with unique features, adhere to the same prompts, demonstrating our model's excellent control over text
conditioning and its ability to preserve individual identities.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/res_main.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</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">Image Generation with text + portrait conditions using ToonYou</h2>
<div class="content has-text-justified">
<p>Given a text prompt and a portrait, our proposed EMMA can integrate with various diffusion models to generate
images in different styles. Here are the images created using EMMA in conjunction with ToonYou.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/res_toonyou.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</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">Image Generation with text + portrait conditions using AnimateDiff</h2>
<div class="content has-text-justified">
<p>Given a portrait and a prompt, our proposed EMMA, combined with the AnimateDiff diffusion model, can generate
images that preserve portrait details while adhering to text instructions.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<div class="gif-row">
<img src="./DreamBooth_files/woman_blue_dress.gif">
<img src="./DreamBooth_files/woman_green_dress.gif">
<img src="./DreamBooth_files/woman_purple_dress.gif">
</div>
</div>
</div>
</div>
</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">Story Telling</h2>
<div class="content has-text-justified">
<p>Images generated by our EMMA with portrait conditions. Two sets of images are generated for two separate stories.
The first set of images is about a mailing woman chased by a dog. The second set of images is about a man finding
treasures.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Teaser. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="content has-text-justified">
<div class="center-image">
<figure>
<img src="./DreamBooth_files/story_diffusion.png" class="interpolation-image"
alt="Interpolate start reference image." />
</figure>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>
@misc{han2024emma,
title={EMMA: Your Text-to-Image Diffusion Model Can Secretly Accept Multi-Modal Prompts},
author={Yucheng Han and Rui Wang and Chi Zhang and Juntao Hu and Pei Cheng and Bin Fu and Hanwang Zhang},
year={2024},
eprint={2406.09162},
archivePrefix={arXiv},
primaryClass={id='cs.CV' full_name='Computer Vision and Pattern Recognition' is_active=True alt_name=None in_archive='cs' is_general=False description='Covers image processing, computer vision, pattern recognition, and scene understanding. Roughly includes material in ACM Subject Classes I.2.10, I.4, and I.5.'}
}
</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link" href="https://arxiv.org/pdf/2406.09162.pdf">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://tencentqqgylab.github.io/EMMA" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This website is 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>
<p>
This means you are free to borrow the <a href="https://github.com/nerfies/nerfies.github.io">source
code</a> of this website,
we just ask that you link back to this page in the footer.
Please remember to remove the analytics code included in the header of the website which
you do not want on your website.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>