-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
354 lines (294 loc) · 14.3 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="LayoutLLM-T2I: Eliciting Layout Guidance from LLM for Text-to-Image Generation">
<meta name="keywords" content="text-to-image generation, Large Language Models, scene synthesis">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>LayoutLLM-T2I</title>
<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="icon" href="./static/images/layout.png">
<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">
<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">LayoutLLM-T2I:</h1>
<h1 class="title is-2 publication-title">Eliciting Layout Guidance from LLM</h1>
<h1 class="title is-2 publication-title" style="margin-top: -17px;">for Text-to-Image Generation</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="#">Leigang Qu</a><sup>1*</sup>,</span>
<span class="author-block">
<a href="https://chocowu.github.io/">Shengqiong Wu</a><sup>1*</sup>,</span>
<span class="author-block">
<a href="https://haofei.vip/">Hao Fei</a><sup>1#</sup>,
</span>
<span class="author-block">
<a href="https://liqiangnie.github.io/">Liqiang Nie</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://www.chuatatseng.com/">Tat-Seng Chua</a><sup>1</sup>,
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><b style="color:#F2A900; font-weight:normal">▶ </b>1. NExT++ Lab, National University of Singapore</span>
<br>
<span class="author-block"><b style="color:#00A4EF; font-weight:normal">▶ </b>2. Harbin Institute of Technology (Shenzhen)</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block" style="font-size: 15px;"><sup>*</sup>Equal Contribution</span>
<span class="author-block" style="font-size: 15px;"><sup>#</sup>Correspondence</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2308.05095.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://huggingface.co/spaces/xxxx/LayoutLLM-T2I" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-laugh"></i>
</span>
<span>Demo</span>
</a>
</span>
<!-- Video Link. -->
<!-- <span class="link-block">
<a href="https://www.youtube.com"
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> -->
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/LayoutLLM-T2I/LayoutLLM-T2I"
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://github.com/LayoutLLM-T2I/LayoutLLM-T2I" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-database"></i>
</span>
<span>Dataset</span>
</a>
</span>
<!-- <span class="link-block">
<a href="https://github.com/LayoutLLM-T2I/LayoutLLM-T2I"
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> -->
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="container is-max-desktop">
<!-- <div class="hero-body">-->
<!-- <h2 class="subtitle has-text-centered">-->
<!-- LayoutLLM for image layout generation based on text inputs. -->
<!-- </h2>-->
<!-- <video id="teaser" autoplay muted loop playsinline height="200%">-->
<!-- <source src="./static/videos/merged_images.mp4"-->
<!-- type="video/mp4">-->
<!-- </video>-->
<!-- </div>-->
<!-- <div class="hero-body">
<h2 class="subtitle has-text-centered">
LayoutGPT for 3D indoor scene synthesis.
</h2>
<video id="teaser" autoplay muted loop playsinline height="100%">
<source src="./static/videos/merged_scenes.mp4"
type="video/mp4">
</video>
</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>
In the text-to-image generation field, recent remarkable progress in Stable Diffusion makes it possible to generate rich kinds of novel photorealistic images.
However, current models still face misalignment issues (<em>e.g.</em>, problematic spatial relation understanding and numeration failure) in complex natural scenes, which impedes the high-faithfulness text-to-image generation.
Although recent efforts have been made to improve controllability by giving fine-grained guidance (<em>e.g.</em>, sketch and scribbles), this issue has not been fundamentally tackled since users have to provide such guidance information manually.
</p>
<p>
In this work, we strive to synthesize high-fidelity images that are semantically aligned with a given textual prompt without any guidance.
Toward this end, we propose a coarse-to-fine paradigm to achieve layout planning and image generation.
Concretely, we first generate the coarse-grained layout conditioned on a given textual prompt via in-context learning based on Large Language Models.
Afterward, we propose a fine-grained object-interaction diffusion method to synthesize high-faithfulness images conditioned on the prompt and the automatically generated layout.
Extensive experiments demonstrate that our proposed method outperforms the state-of-the-art models in terms of cross-modal text-layout alignment and high-faithfulness image generation.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
<!-- 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://www.youtube.com/"
frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
</div>
</div>
</div> -->
<!--/ Paper video. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Motivation</h2>
<div class="content has-text-justified">
<p>
Despite the satisfactory performance achieved by recent SD-based models, synthesizing high-faithful images in complex scenes is still challenging, such as <font color="#4C76C6"><em>problematic spatial relation understanding</em></font> and <font color="#4C76C6"><em>numeration failure</em></font>, as shown in the following figure.
However, to achieve high-faithfulness image synthesis, the existing models suffer from the following challenges:
<ol>
<li>
<b>Layout Planning</b> requires abstract spatial imagination and analysis capabilities. The limited annotated layout data and intrinsic inductive bias make it difficult for existing diffusion methods to accurately and aesthetically generate layouts. Although notable efforts have been dedicated to synthesizing complex scenes by manually providing guidance information, these strategies suffer from weak flexibility and low efficiency since they heavily rely on extra labor-intensive guidance.
</li>
<li>
<b>Relation Modeling</b>, <em>e.g.</em>, the high-level spatial and semantic relations, plays a pivotal role in understanding, imagining, and depicting complex scenes for T2I models, but it is still under-explored owing to the complex and ever-changing environments in real life.
</li>
</ol>
</p>
<img class="columns is-centered has-text-centered" src="./static/images/intro.jpg" alt="Teaser" width="60%" style="margin:0 auto">
</div>
<br/>
</div>
</div>
<!-- Framework -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Framework</h2>
<div class="content has-text-justified">
<p>
In Figure 2, we illustrate the overall architecture of the proposed layout-guided diffusion model, consisting of two modules:
<ol>
<li>
<b>Text-to-layout induction module</b> infers a coarse-grained layout via an LLM conditioned the given textual prompt.
</li>
<li>
<b>Layout-guided image generation module</b> synthesizes the final image based on the prompt and the generated layout, .
</li>
</ol>
</p>
<img class="columns is-centered has-text-centered" src="./static/images/framework.png" alt="Teaser" width="100%" style="margin:0 auto">
</div>
<br/>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Generated Images</h2>
<h3 class="title is-4">Examples: Spatial, Semantic</h3>
<div class="content has-text-justified">
<p>
Results on the <strong>Spatial</strong>, and <strong>Semantic</strong> inputs.
The ground-truth (GT) image, the ground-truth layout with the generated image (GT*), the layout generated from LayoutDM, and our results for each prompt are shown from left to right.
</p>
<img src="./static/images/case-4.jpg" alt="Teaser" width="100%">
</div>
<h3 class="title is-4">Examples: Numerical, Mixed, Null</h3>
<div class="content has-text-justified">
<p>
Results on the <strong>Numerical</strong>, <strong>Mixed</strong>, and <strong>Null</strong> inputs.
</p>
<img src="./static/images/case-3.jpg" alt="Teaser" width="100%">
</div>
<br/>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Related Links</h2>
<div class="content has-text-justified">
<p>
You may refer to previous works such as <a href="https://gligen.github.io/">GLIGEN</a>, <a href="https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/text2img">Stable Diffusion</a>, and <a href="https://platform.openai.com/docs/models/gpt-3-5">GPT3.5</a>, which serve as foundational frameworks for our LayoutLLM-T2I framework and code repository.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@article{qu2023layoutllm,
title={LayoutLLM-T2I: Eliciting Layout Guidance from LLM for Text-to-Image Generation},
author={Leigang Qu, Shengqiong Wu, Hao Fei, Liqiang Nie, Tat-Seng Chua},
journal={Proceedings of the {ACM} International Conference on Multimedia},
year={2023}
}
</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p style="text-align: center;">
The webpage is built based on <a href="https://github.com/nerfies/nerfies.github.io">Nerfies</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>