/
index.html
365 lines (328 loc) · 16.4 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="DESCRIPTION META TAG">
<meta property="og:title" content="SOCIAL MEDIA TITLE TAG"/>
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG"/>
<meta property="og:url" content="URL OF THE WEBSITE"/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
<meta property="og:image" content="static/image/your_banner_image.png" />
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="630"/>
<meta name="twitter:title" content="TWITTER BANNER TITLE META TAG">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<meta name="twitter:image" content="static/images/your_twitter_banner_image.png">
<meta name="twitter:card" content="summary_large_image">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords" content="KEYWORDS SHOULD BE PLACED HERE">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Towards Open-ended Visual Quality Comparison</title>
<link rel="icon" type="image/x-icon" href="static/images/logo.pic.jpg">
<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">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.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>
<script type="module" src="https://gradio.s3-us-west-2.amazonaws.com/4.12.0/gradio.js"></script>
<style>
.darkblue {
color: darkblue;
}
pre {
white-space: pre-wrap; /* Since CSS might be used, ensure that white-space is preserved */
}
</style>
</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">Towards Open-ended Visual Quality Comparison</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://teowu.github.io" target="_blank">Haoning Wu</a><sup>1*</sup>,
</span>
<span class="author-block">
<a href="https://github.com/h4nwei" target="_blank">Hanwei Zhu</a><sup>2*</sup>,
</span>
<span class="author-block">
<a href="https://github.com/zzc-1998" target="_blank">Zicheng Zhang</a><sup>3*</sup>,
</span>
<span class="author-block">
<a href="https://github.com/ZhangErliCarl/" target="_blank">Erli Zhang</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://chaofengc.github.io" target="_blank">Chaofeng Chen</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://liaoliang92.github.io" target="_blank">Liang Liao</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://github.com/lcysyzxdxc" target="_blank">Chunyi Li</a><sup>3</sup>,
</span>
<span class="author-block">
<a href="https://github.com/AnnanWangDaniel" target="_blank">Annan Wang</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://wenxiusun.com" target="_blank">Wenxiu Sun</a><sup>4</sup>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=uT9CtPYAAAAJ&hl=en" target="_blank">Qiong Yan</a><sup>4</sup>,
</span>
<span class="author-block">
<a href="https://jhc.sjtu.edu.cn/~xiaohongliu/" target="_blank">Xiaohong Liu</a><sup>3</sup>,
</span>
<span class="author-block">
<a href="https://ee.sjtu.edu.cn/en/FacultyDetail.aspx?id=24&infoid=153&flag=153" target="_blank">Guangtao Zhai</a><sup>3</sup>,
</span>
<span class="author-block">
<a href="https://www.cs.cityu.edu.hk/~shiqwang/" target="_blank">Shiqi Wang</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://personal.ntu.edu.sg/wslin/Home.html" target="_blank">Weisi Lin</a><sup>1</sup>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>Nanyang Technological University</span>
<span class="author-block"><sup>2</sup>City University of Hong Kong</span>
<span class="author-block"><sup>3</sup>Shanghai Jiao Tong University</span>
<span class="author-block"><sup>4</sup>Sensetime Research</span>
<span class="eql-cntrb"><small><br><sup>*</sup>Equal Contribution.</small></span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- Arxiv PDF link -->
<span class="link-block">
<a href="https://arxiv.org/abs/2402.16641" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>ArXiv (Abstract)</span>
</a>
</span>
<!-- huggingface -->
<span class="link-block">
<a href="https://huggingface.co/q-future/co-instruct" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-smile"></i>
</span>
<span>Co-Instruct (Model)</span>
</a>
</span>
<!-- Github link -->
<span class="link-block">
<a href="https://github.com/Q-Future/Co-Instruct" 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>
<!-- huggingface -->
<span class="link-block">
<a href="https://huggingface.co/spaces/q-future/Co-Instruct" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-smile"></i>
</span>
<span>HF Demo</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section" style="background-color:#efeff081">
<div class="container is-max-desktop" id="gradio">
<gradio-app src="https://q-future-co-instruct.hf.space/"></gradio-app>
</div>
</section>
<!-- Paper abstract -->
<section class="section hero is-light">
<div class="container is-max-desktop">
<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>
Comparative settings (e.g., pairwise choice, listwise ranking) have been adopted by a wide range of subjective studies for image quality assessment (IQA), as it inherently standardizes the evaluation criteria across different observers and offers more clear-cut responses. In this work, we extend the edge of emerging large multi-modality models (LMMs) to further advance visual quality comparison into open-ended settings, that 1) can respond to open-range questions on quality comparison; 2) can provide detailed reasonings beyond direct answers. To this end, we propose the Co-Instruct. To train this first-of-its-kind open-source open-ended visual quality comparer, we collect the Co-Instruct-562K dataset, from two sources: (a) LMM-merged single image quality description, (b) GPT-4V "teacher" responses on unlabeled data. Furthermore, to better evaluate this setting, we propose the MICBench, the first benchmark on multi-image comparison for LMMs. We demonstrate that Co-Instruct not only achieves 30% higher superior accuracy than state-of-the-art open-source LMMs but also outperforms GPT-4V (its teacher) on both existing related benchmarks and the proposed MICBench."
</p>
</div>
<img src="static/images/teaser.png" , width="800" />
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- Teaser -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="columns is-centered has-text-centered">
<div class="column is-three-fifths">
<h5 class="title">Motivation</h5>
<div class="content has-text-justified">
<p>
The motivation of open-ended visual quality comparison: comparative settings can effectively avoid the <strong>ambiguity on absolute evaluations</strong> for single images, and provide more clear-cut judgements to serve as downstream guidances.
</p>
</div>
<img src="static/images/motivation.png" , width="1400" />
</div>
</div>
</div>
</section>
<!-- Teaser -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="columns is-centered has-text-centered">
<div class="column is-three-fifths">
<h5 class="title">Dataset: Co-Instruct-562K</strong></h5>
<div class="content has-text-justified">
<p>
We adopt two non-perfect supervisors: <em>(1)</em> <strong>Merge2Compare:</strong> Originated from single image human quality descriptions of 19K images in Q-Pathway, we randomly match them into 100K groups,
removing the most similar descriptions using a text embedding model. Similar to the construction of LLaVA-150K, we prompt a single-modal LLM to compare multiple human descriptions in a group,
and merge them into 100K pseudo comparisons. <em>(2)</em> <strong>Teach2Compare:</strong> Observing that GPT-4V has especially high accuracy on pairwise settings among existing LMMs,
we leverage GPT-4V responses to expand our dataset. We collect 9K unlabeled images and match them into 30K image groups (2-4 images per group) and obtain GPT-4V responses on both caption-like general comparisons
and question-answer pairs for comparisons. By integrating Q-Instruct-200K (on single images), Merge2Compare, and Teach2Compare we construct the <strong>Co-Instruct-562K</strong>,
the first instruction tuning dataset designed for open-ended multi-image quality comparison.
</p>
</div>
<img src="static/images/dataset.png" , width="1400" />
</div>
</div>
</div>
</section>
<!-- Teaser -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="columns is-centered has-text-centered">
<div class="column is-three-fifths">
<h5 class="title">Structure</h5>
<div class="content has-text-justified">
<p>
<p>To correctly refer to each specific image during conversation, we define a specific image-text interleaved format to handle multi-image cases, as follows:</p>
<pre>
User: The first image: <span class="darkblue"><img<sub>0</sub>></span> The second image: <span class="darkblue"><img<sub>1</sub>></span> ... <query>
Assistant: <response>
</pre>
Moreover, as we need to feed multiple images together during instruction tuning, adopting the most popular LLaVA structure that linearly projects visual embeddings will exceed the context window of the language models and cause errors. Henceforth, we adopt an alternative visual abstractor structure to first reduce visual token length (from 1025 to 65 tokens per image), and then concatenate them with text embeddings to pass to language decoders.
</p>
</div>
<img src="static/images/framework.png" , width="900" />
</div>
</div>
</div>
</section>
<!-- Teaser -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="columns is-centered has-text-centered">
<div class="column is-three-fifths">
<h5 class="title">MICBench</h5>
<div class="content has-text-justified">
<p>
We build the <strong>MICBench</strong> to cover the open-ended evaluation settings on <strong>groups of three or four images</strong>, as a complementary of existing evaluation settings.
It contains 2,000 groups of open-range questions equipped with multiple candidates, including Sourcing Diverse Image Groups and Multi-choice Questions (MCQs).
</p>
</div>
<img src="static/images/micbench.png" , width="900" />
</div>
</div>
</div>
</section>
<!-- Image carousel -->
<section class="hero is-small">
<div class="hero-body">
<div class="container">
<div id="results-carousel" class="carousel results-carousel">
<div class="item">
<!-- Your image here -->
<img src="static/images/Q-A1.png" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Results on <span style="font-style: italic;">Q-Bench<sup style="font-family: monospace;">PAIR</sup>-A1</span>.
</h2>
</div>
<div class="item">
<!-- Your image here -->
<img src="static/images/Q-A2.png" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Results on <span style="font-style: italic;">Q-Bench<sup style="font-family: monospace;">PAIR</sup>-A2</span>.
</h2>
</div>
<div class="item">
<!-- Your image here -->
<img src="static/images/2afc.png" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Results on 2AFC-LMM.
</h2>
</div>
<div class="item">
<!-- Your image here -->
<img src="static/images/micbench-result.png" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Results of MICBench.
</h2>
</div>
</div>
</div>
</div>
</section>
<!-- End image carousel -->
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@misc{wu2024openended,
title={Towards Open-ended Visual Quality Comparison},
author={Haoning Wu and Hanwei Zhu and Zicheng Zhang and Erli Zhang and Chaofeng Chen and Liang Liao and Chunyi Li and Annan Wang and Wenxiu Sun and Qiong Yan and Xiaohong Liu and Guangtao Zhai and Shiqi Wang and Weisi Lin},
year={2024},
eprint={2402.16641},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
</code></pre>
</div>
</section>
<!--End BibTex citation -->
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This page was built using the <a href="https://github.com/eliahuhorwitz/Academic-project-page-template" target="_blank">Academic Project Page Template</a>.
<br> This website is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
<!-- Statcounter tracking code -->
<!-- You can add a tracker to track page visits by creating an account at statcounter.com -->
<!-- End of Statcounter Code -->
</body>
</html>