-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
380 lines (333 loc) · 20.6 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriately as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="See, Say, and Segment: Teaching LMMs to Overcome False Premises">
<meta property="og:title" content="SESAME" />
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG" />
<meta property="og:url" content="http://see-say-segment.github.io" />
<!-- Path to banner image, should be in the path listed below. Optimal dimensions are 1200X630-->
<meta property="og:image" content="/static/images/sesame_banner.png" />
<meta property="og:image:width" content="1200" />
<meta property="og:image:height" content="630" />
<meta name="twitter:title" content="See, Say, and Segment: Teaching LMMs to Overcome False Premises">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimensions are 1200X600-->
<meta name="twitter:image" content="static/images/sesame_twitter_banner.png">
<meta name="twitter:card" content="SESAME Project Logo: A cartoon bun with sesame seeds on top.">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords"
content="Large Multimodal Models, segmentation, reasoning, VQA, referring expression segmentation">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>See, Say and Segment: Teaching LMMs to Overcome False Premises</title>
<link rel="icon" type="image/x-icon" href="static/images/favicon_io/favicon.ico">
<link rel="apple-touch-icon" sizes="180x180" href="static/images/favicon_io">
<link rel="icon" type="image/png" sizes="32x32" href="static/images/favicon_io/favicon-32x32.png">
<link rel="icon" type="image/png" sizes="16x16" href="static/images/favicon_io/favicon-16x16.png">
<link rel="manifest" href="static/images/favicon_io/site.webmanifest">
<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/fontawesome/css/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/fontawesome/js/fontawesome.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 is-vcentered">
<div class="column"><img src="static/images/sesame_logo.png" height="187" width="187"></div>
<div class="column has-text-centered is-four-fifths is-vcentered">
<h1 class="title is-1 publication-title">See, Say, and Segment: Teaching LMMs to Overcome False
Premises</h1>
</div>
</div>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered is-four-fifths">
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://tsunghan-wu.github.io/" target="_blank">Tsung-Han
Wu</a><sup>*</sup>,</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=s0Fof5IAAAAJ" target="_blank">Giscard
Biamby</a><sup>*</sup>,</span>
<span class="author-block">
<a href="https://dchan.cc/" target="_blank">David Chan</a>,
</span>
<span class="author-block">
<a href="https://www.lisabdunlap.com/" target="_blank">Lisa Dunlap</a>,
</span>
<span class="author-block">
<a href="https://ritwikgupta.me/" target="_blank">Ritwik Gupta</a>,
</span>
<span class="author-block">
<a href="http://people.eecs.berkeley.edu/~xdwang/" target="_blank">Xudong Wang</a>,
</span>
<span class="author-block">
<a href="https://people.eecs.berkeley.edu/~jegonzal/" target="_blank">Joseph E. Gonzalez</a>,
</span>
<span class="author-block">
<a href="https://people.eecs.berkeley.edu/~trevor/" target="_blank">Trevor Darrell</a>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">UC Berkeley<br><span
class="author-block" style="font-weight: bold;">CVPR 2024</span></span>
<span class="eql-cntrb"><small><br><sup>*</sup>Indicates Equal
Contribution</small></span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a href="https://arxiv.org/abs/2312.08366" 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://youtu.be/-TXCR-m3MJ4?si=I2FxhXdsn860aZxC" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa-brands fa-youtube"></i>
</span>
<span>Video</span>
</a>
</span>
<span class="link-block">
<a href="https://huggingface.co/collections/tsunghanwu/sesame-666c9b43efff2acaafc61882" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon" style="vertical-align: middle; font-size: 20px;">🤗</span>
<span>Model</span>
</a>
</span>
<span class="link-block">
<a href="https://drive.google.com/file/d/1mA3kcY3QiAZz1Zr89MCKYd7e3LBIwUzl/view?usp=drive_link" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-database"></i>
</span>
<span>Dataset</span>
</a>
</span>
<span class="link-block">
<a href="https://github.com/see-say-segment/sesame" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa-brands fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Paper abstract -->
<section class="section hero ">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">How do LMMs handle false premise segmentation queries?</h2>
<div class="content has-text-justified">
<p>
Contemporary open-source Large MultiModal Models (LMMs) combined with segmentation decoders (such as <a href="https://github.com/dvlab-research/LISA?tab=readme-ov-file">LISA</a>) are able to generate awesome segmentation masks but have difficulty on expressions which refer to something that is not present in the image. <strong>SESAME</strong>, our <strong>SE</strong>e-<strong>SA</strong>y-seg<strong>ME</strong>nt LMM, uses joint training to overcome this problem.
</p>
</div>
<img src="static/images/figures/fig1_SSS.jpg"
alt="Figure 1: False Premise Referring Expression Segmentation Examples with prior work and our method." />
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- SESAME -->
<section class="section ">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">Advanced Problem Setting: See, Say and Segment</h2>
</div>
</div>
<!-- <div id="results-carousel" class="carousel results-carousel"> -->
<div class="content has-text-justified">
<p>
We introduce <strong>a Novel Problem Setting</strong>, requiring LMMs that can See, Say and Segment. Specifically, we require these models to
<ol>
<li><b>See</b> by detecting if an object from the query is present in an image,</li>
<li><b>Say</b> something about the object itself if it’s not there and suggest alternatives to the user’s query,</li>
<li><b>Segment</b> by showing where in an image an existent object is grounded.</li>
</ol>
</p>
</div>
<div class="item is-vcentered">
<img src="static/images/figures/fig2_SSS.gif" alt="Figure 2: Diagram of SESAME Model Framework" />
</div>
<!-- </div> -->
</div>
</section>
<!-- DATASET -->
<section class="section hero ">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">FP-RefCOCO: A Novel Dataset</h2>
<div class="content has-text-justified">
<p>
To facilitate training and evaluation of this new class of models, we introduce <strong>a new dataset and benchmark, FP-RefCOCO, FP-RefCOCO+ and FP-RefCOCOg</strong>.
</p>
<!-- <img class="img" src="static/images/figures/fig3_SSS.png"> -->
</div>
<div class="content has-text-justified">
<!-- <br /> -->
Using refCOCO for base images, we employ an LLM to augment a false-premise referring segmentation dataset with <strong>context-aware false premise queries: similar objects, attributes, and relations.</strong> Although existing datasets such as R-RefCOCO(+/g) also include queries referring to non-existent items in images, their method of generating negative expressions through naive random sampling often lacks context-awareness. This limitation significantly reduces their effectiveness for false-premise correction tasks.
</div>
<div class="item">
<img src="static/images/figures/data_1.png" alt="MY ALT TEXT" />
</div>
<div class="content has-text-justified">
<br/>
Consider an image with a cat on a chair: contextually valid false premises that could be logically corrected to "a cat on the chair" might include phrases like "a cat under the chair" or "a dog on the chair." However, prior datasets such as R-RefCOCO typically produces less suitable examples, such as ``a pizza on the chair” or "a cat behind the people," which do not align with realistic model correction expectations.
</div>
</div>
</div>
</section>
<section class="section hero">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">Visualization Results: Referring Segmentation</h2>
<div class="content has-text-justified">
<p>
When SESAME is presented with false premises queries with similar objects, attributes, concepts, or activities, it can not only deny these queries but also uses commonsense reasoning to propose relevant alternatives that align with human understanding. For queries that are entirely irrelevant, SESAME will simply reject them without generating any baseless or speculative results.
</p>
<!-- <img class="img" src="static/images/figures/fig3_SSS.png"> -->
</div>
</div>
</div>
<div id="results-carousel" class="carousel results-carousel">
<div class="item">
<img src="static/images/figures/ref_exp1.jpg"
alt="Example 1." />
</div>
<div class="item">
<img src="static/images/figures/ref_exp2.jpg"
alt="Example 2." />
</div>
<div class="item">
<img src="static/images/figures/ref_exp3.jpg"
alt="Example 3." />
</div>
</div>
<br/>
<br/>
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">Visualization Results: Reasoning Segmentation</h2>
<div class="content has-text-justified">
<p>
SESAME is adaptable for complex "reasoning segmentation" tasks, where objects are implicitly implied instead of directly mentioned. By training with specially curated data for false-premise reasoning segmentation, our model can not only dismiss incorrect queries but also optionally suggest an alternative similar concept.
</p>
<!-- <img class="img" src="static/images/figures/fig3_SSS.png"> -->
</div>
</div>
</div>
<div id="results-carousel" class="carousel results-carousel">
<div class="item">
<img src="static/images/figures/reason_exp1.png"
alt="Example 1." />
</div>
<div class="item">
<img src="static/images/figures/reason_exp2.png"
alt="Example 2." />
</div>
<div class="item">
<img src="static/images/figures/reason_exp3.png"
alt="Example 3." />
</div>
</div>
<br/>
<br/>
<div class="columns is-centered has-text-centered">
<div class="column is-full">
<h2 class="title is-4">Visualization Results: Ability to handle complex instructions</h2>
<div class="content has-text-justified">
<p>
SESAME stands out by processing complex input instructions, including segmenting alternate objects based on conditional queries and conducting basic Visual Question Answering (VQA) without producing segmentation masks. This versatility, unlike prior models like LISA, opens the door for more human-like interactions and the possibility of extending SESAME to multi-round interactions.
</p>
<div class="item">
<img src="static/images/figures/multiround.jpg" alt="MY ALT TEXT" />
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Paper poster -->
<!-- <section class="hero is-small is-light">
<div class="hero-body">
<div class="container">
<h2 class="title">Poster</h2>
<iframe src="static/pdfs/sample.pdf" width="100%" height="550">
</iframe>
</div>
</div>
</section> -->
<!--End paper poster -->
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@inproceedings{wu2024see,
title={See Say and Segment: Teaching LMMs to Overcome False Premises},
author={Wu, Tsung-Han and Biamby, Giscard and Chan, David and Dunlap, Lisa and Gupta, Ritwik and Wang, Xudong and Gonzalez, Joseph E and Darrell, Trevor},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={13459--13469},
year={2024}
}</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> which was adopted from the <a
href="https://nerfies.github.io" target="_blank">Nerfies</a> project page.
You are free to borrow the of this website, we just ask that you link back to this page in
the footer. <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>