-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
520 lines (456 loc) · 22 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- <meta name="description"
content="HyperNeRF handles topological variations by modeling a family of shapes in a higher-dimensional space, thereby producing more realistic renderings and more accurate geometric reconstructions.">
<meta name="keywords" content="HyperNeRF, Nerfies, D-NeRF, NeRF">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta property="og:image" content="./static/images/thumbnail.png"/>
<link rel="image_src" href="./static/images/thumbnail.png">
<link rel="icon"
type="image/x-icon"
href="./static/images/favicon.ico"/> -->
<title>RobustSAM: Segment Anything Robustly on Degraded Images
</title>
<style>
.img-container {
text-align: center; /* Center the text and images */
margin-bottom: 20px; /* Adds some space between the image groups */
}
.image-with-caption {
display: inline-block; /* Allows multiple elements side by side */
margin: 10px; /* Adds some space around each image and caption */
}
.custom-gif {
width: 200px; /* Adjust based on your preference */
height: auto; /* Maintain aspect ratio */
margin-top: 10px; /* Space between caption and image */
}
.caption {
font-weight: bold; /* Optional: makes the caption text bold */
}
</style>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-EDF010G6PN"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-EDF010G6PN');
</script>
<script type="module"
src="https://unpkg.com/@google/model-viewer/dist/model-viewer.min.js"></script>
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.6.0/jquery.min.js"></script>
<script type="text/javascript" src="https://code.jquery.com/jquery-1.11.0.min.js"></script>
<script type="text/javascript" src="https://code.jquery.com/jquery-migrate-1.2.1.min.js"></script>
<script src="https://unpkg.com/interactjs/dist/interact.min.js"></script>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" type="text/css" href="./static/slick/slick.css"/>
<link rel="stylesheet" type="text/css" href="./static/slick/slick-theme.css"/>
<link rel="stylesheet" href="./static/css/bulma.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 defer src="./static/js/fontawesome.all.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">
<!-- <div class="columns is-centered">
<div class="column is-4 has-text-centered">
<img src="static/images/logo.svg" alt="HyperNeRF"/>
</div>
</div> -->
<div class="container has-text-centered">
<h1 class="title is-1 publication-title">
RobustSAM: Segment Anything Robustly on Degraded Images
</h1>
<div class="is-size-5 publication-authors">
<div class="author-block"><a href="https://sites.google.com/view/weitingchen/home">Wei-Ting Chen</a> <sup> 1, 2</sup></div>
<div class="author-block"><a href="https://scholar.google.com/citations?user=lnjDFYEAAAAJ&hl=en"> Yu-Jiet Vong</a> <sup> 1</sup></div>
<div class="author-block"><a href="https://homepage.ntu.edu.tw/~sykuo"> Sy-Yen Kuo</a> <sup> 1</sup></div>
<div class="author-block"><a href="https://sizhuoma.netlify.app/"> Sizhuo Ma</a> <sup> 2*</sup></div>
<div class="author-block"><a href="https://jianwang-cmu.github.io/"> Jian Wang</a> <sup> 2*</sup></div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>National Taiwan University,</span>
<span class="author-block"><sup>2</sup>Snap Inc.</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"> </sup>* Co-corresponding authors</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"> </sup>(CVPR 2024 Highlight)</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://openaccess.thecvf.com/content/CVPR2024/html/Chen_RobustSAM_Segment_Anything_Robustly_on_Degraded_Images_CVPR_2024_paper.html"
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/pdf/2406.09627"
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/robustsam/RobustSAM"
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>
<!-- Dataset Link. -->
<span class="link-block">
<a href="https://github.com/robustsam/RobustSAM"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="far fa-images"></i>
</span>
<span>Data</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="hero-body">
<div class="container is-max-desktop">
<!-- <video id="teaser" autoplay controls muted loop playsinline height="100%">
<source src="./static/images/teaser.mp4"
type="video/mp4">
</video> -->
<!-- <h2 class="subtitle has-text-centered"> -->
<!-- <i>RobustSAM</i> outperfroms SAM in various degradations. -->
<!-- <i>RobustSAM</i> handles topological variations by modeling a
family of shapes in a higher-dimensional space, thereby producing more realistic renderings
and more accurate geometric reconstructions. -->
<!-- </h2> -->
</div>
</div>
</section>
<div class="img-container">
<div class="image-with-caption">
<div class="caption">Blur</div>
<img class="custom-gif" src="./robustsam_file/gif_output/blur_back_n_forth.gif">
</div>
<div class="image-with-caption">
<div class="caption">Haze</div>
<img class="custom-gif" src="./robustsam_file/gif_output/haze_back_n_forth.gif">
</div>
<div class="image-with-caption">
<div class="caption">Lowlight</div>
<img class="custom-gif" src="./robustsam_file/gif_output/lowlight_back_n_forth.gif">
</div>
<div class="image-with-caption">
<div class="caption">Rain</div>
<img class="custom-gif" src="./robustsam_file/gif_output/rain_back_n_forth.gif">
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column is-three-quarters">
<p>
<b>RobustSAM</b> </i> outperforms with precise boundaries and
intact structures, where SAM falters with errors and fragmentation.
Red star points and bounding boxes are our examples' input prompts.
</p>
</div>
</div>
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Centered Responsive YouTube Video Embed</title>
<style>
.video-container {
display: flex;
justify-content: center;
align-items: center;
width: 100%;
}
.video-container iframe {
width: 100%;
max-width: 720px; /* Optional: Set a max-width if you want */
height: auto;
aspect-ratio: 16 / 9;
}
</style>
</head>
<body>
<div class="video-container">
<iframe src="https://www.youtube.com/embed/Awukqkbs6zM?si=vQEW91tN4j_iRJNs"
title="YouTube video player" frameborder="0"
allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share"
referrerpolicy="strict-origin-when-cross-origin"
allowfullscreen>
</iframe>
</div>
</body>
</html>
<!-- <iframe width="560" height="315"
src="https://www.youtube.com/embed/Awukqkbs6zM?si=vQEW91tN4j_iRJNs"
title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share"
referrerpolicy="strict-origin-when-cross-origin"
allowfullscreen>
</iframe> -->
<!-- </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>
Segment Anything Model (SAM) has emerged as a transformative approach in image segmentation,
acclaimed for its robust zero-shot segmentation capabilities and flexible prompting system.
Nonetheless, its performance is challenged by images with degraded quality. Addressing this
limitation, we propose the Robust Segment Anything Model (RobustSAM), which enhances SAM's
performance on low-quality images while preserving its promptability and zero-shot generalization.
</p>
<p>
Our method leverages the pre-trained SAM model with only marginal parameter increments and computational requirements.
The additional parameters of RobustSAM can be optimized within 30 hours on eight GPUs, demonstrating
its feasibility and practicality for typical research laboratories. We also introduce the Robust-Seg dataset,
a collection of 688K image-mask pairs with different degradations designed to train and evaluate our model optimally.
Extensive experiments across various segmentation tasks and datasets confirm RobustSAM's superior performance,
especially under zero-shot conditions, underscoring its potential for extensive real-world application. Additionally,
our method has been shown to effectively improve the performance of SAM-based downstream tasks such as single
image dehazing and deblurring.
</p>
<!-- <p>
We address this limitation by lifting
NeRFs into a higher dimensional space, and by representing the 5D radiance field
corresponding to each individual input image as a slice through this "hyper-space". Our
method is inspired by level set methods, which model the evolution of surfaces as slices
through a higher dimensional surface. We evaluate our method on two tasks: (i)
interpolating smoothly between "moments", i.e., configurations of the scene, seen in the
input images while maintaining visual plausibility, and (ii) novel-view synthesis at
fixed moments. We show that our method, which we dub HyperNeRF, outperforms existing
methods on both tasks. Compared to Nerfies, <i>HyperNeRF</i> reduces average error rates by
4.1% for interpolation and 8.6% for novel-view synthesis, as measured by LPIPS.
</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-2">Video</h2>
<div class="publication-video">
<iframe width="640" height="480" src="https://www.youtube.com/watch?v=bx0He5eE8fE"
title="YouTube video player" frameborder="0"
allow="accelerometer; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
allowfullscreen></iframe>
</div>
</div>
</div> -->
<!--/ Paper video. -->
</div>
</section>
<hr/>
<section class="section">
<!-- <div class="container is-max-desktop">
<h2 class="title is-3">Motivation</h2>
<div class="columns is-centered">
<!-- Level Sets. -->
<!-- <div class="column is-half">
<div class="content">
<h3 class="title is-4">Level-Set Methods</h3>
<div class="level-set has-text-justified">
<p>
HyperNeRF represents changes in scene topology by providing a
NeRF with a higher-dimensional input. This is inspired by level-set methods.
Level-set methods provide a means to model a family of topologically-varying shapes as
slices of a higher dimensional auxiliary function. For example, these shapes
</p>
<div class="level-set-shapes">
<img src="./static/figures/level_set/0.svg"/>
<img src="./static/figures/level_set/1.svg"/>
<img src="./static/figures/level_set/2.svg"/>
</div>
<p>
can be represented as slices through this auxiliary shape
</p>
<model-viewer class="level-set-slices"
src="./static/figures/level_set/level_set_3d.glb"
alt="Slices through a 3D ambient surface."
environment-image="neutral" auto-rotate camera-controls></model-viewer>
<p>
We can naturally model topologically varying shapes by just moving the
slicing plane along the higher dimensions. For example, this animation was generated
by
moving the slicing plane from top to bottom:
</p>
<div class="has-text-centered">
<video class="level-set-interpolate" controls autoplay loop playsinline muted height="100%">
<source src="./static/figures/level_set/interpolate2.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div> -->
<!--/ Level Sets. -->
<!-- <div class="column is-half">
<div class="content has-text-justified">
<h3 class="is-4">Slicing Surfaces</h3>
<p>
Consider the follow shapes, which have different permutations of O xand X.
</p>
<div class="level-set-ox-shapes">
<img src="./static/figures/level_set/ox/0.svg"/>
<img src="./static/figures/level_set/ox/1.svg"/>
<img src="./static/figures/level_set/ox/2.svg"/>
<img src="./static/figures/level_set/ox/3.svg"/>
</div>
<p>
Traditionally, level-set methods use straight planes to slice the higher-dimensional
surface:
</p>
<model-viewer class="level-set-slices"
src="./static/figures/level_set/ox/ox_ap.glb"
alt="Slices through a 3D ambient surface."
environment-image="neutral" auto-rotate camera-controls></model-viewer>
<p>
This means the higher-dimensional shape must contain copies of the same shape since
each permutation has to lie along a single straight slice through the z-axis. If we let
the slicing plane bend, it results in a much cleaner template:
</p>
<model-viewer class="level-set-slices"
src="./static/figures/level_set/ox/ox_ds.glb"
alt="Slices through a 3D ambient surface."
environment-image="neutral" auto-rotate camera-controls></model-viewer>
<p>Please see the paper for details.</p>
</div>
</div>
</div>
</div> --> -->
<div class="container is-max-desktop">
<div class="content has-text-justified">
<h2 class="title is-3">Architecture of RobustSAM</h2>
<p>
The key contribution of RobustSAM
is the <b>Anti-Degradation Output Token Generation (AOTG)</b> and <b>Anti-Degradation Mask Feature Generation
(AMFG)</b> modules, which extract degradation-invariant information that is aligned those extracted from
clear images by the original SAM.
<div class="has-text-centered">
<!-- <embed src="./robustsam_file/architecture.pdf" width="800px" height="2100px" /> -->
<img style="width: 100%;" src="./robustsam_file/architecture.jpg"
alt="RobustSAM architecture."/>
<div class="content has-text-justified">
<p>
RobustSAM augments the original SAM by incorporating five essential components
(in purple). During training, clear images are fed through the original SAM modules (in gray) to produce features for clear scenes.
Subsequently, degraded images, generated through augmentation of clear inputs, are processed by RobustSAM, yielding features for
degraded scenarios. These are then refined via Anti-degradation modules, ensuring consistency with features from clear scenes. This
methodology, supported by a segmentation loss, achieves precise segmentation outcomes in both clear and degraded image conditions.
During inference, only RobustSAM is used to predict a segmentation mask from an input image. Note: The prompt encoder is excluded for
conciseness, and the padlock icons represent fixed components loaded from the original SAM model that are not updated during training.
</div>
</div>
<h3 class="title is-4">Overview of the proposed AMFG and AOTG modules</h3>
<!-- <embed src="./robustsam_file/style-suppresion.pdf" width="800px" height="2100px" /> -->
<img style="width: 100%;" src="./robustsam_file/module.jpg"
alt="AMFG and AOTG modules."/>
<div class="content has-text-justified">
<p>
The Anti-Degradation Mask Feature Generation and Output Token Generation modules significantly enhance image quality and segmentation by normalizing degradation effects while preserving essential content. Utilizing Instance and Batch Normalization, the model stabilizes content under various conditions. A Fourier Degradation Suppression module further isolates degradation, focusing on maintaining structural integrity. These modules ensure the generation of robust, degradation-invariant features for precise image assessment and segmentation.
</div>
<!-- <p>
HyperNeRF leverages main idea of level set methods by using a template NeRF which lives in
higher dimensions. In addition to the spatial coordinates (X, Y, Z), the NeRF MLP takes
additional higher dimensional coordinates W<sub>1</sub> and W<sub>2</sub>. We call these
the "ambient dimensions".
</p>
<p>
Here is an interactive viewer for the hyper-space of capture shown in the teaser. Drag the
<span style="color: #29e">blue cursor</span> around to change the ambient dimension rendered
on the right.
</p> -->
</div>
<!-- <div class="columns is-centered">
<div class="column is-half">
<div class="hyper-space-wrapper has-text-centered">
<div class="hyper-space-axis">
<div class="hyper-space">
<div class="hyper-space-cursor"></div>
</div>
</div>
Ambient Dimension Coordinates
<br/>
<small>(Background shows log density of coordinate)</small>
</div>
</div>
<div class="column is-half has-text-centered">
<div class="hyper-grid-wrapper">
<div class="hyper-grid-rgb">
<img src="./static/figures/hyper_grid.jpg"/>
</div>
</div>
The hyper-space template rendered from a fixed viewpoint.
</div>
</div> -->
</div>
</section>
<hr/>
<section class="section" id="BibTeX">
<div class="container content is-max-desktop">
<h2 class="title">BibTeX</h2>
<pre><code>@inproceedings{chen2024robustsam,
author = {Chen, Wei-Ting and Vong, Yu-Jiet and Kuo, Sy-Yen and Ma, Sizhou and Wang, Jian},
title = {RobustSAM: Segment Anything Robustly on Degraded Images},
journal = {CVPR},
issue_date = {2024}
}</code></pre>
</div>
</section>
<!-- <section class="section" id="acknowledgements">
<div class="container content is-max-desktop">
<h2 class="title">Acknowledgements</h2>
<p>Special thanks to <a href="https://homes.cs.washington.edu/~holynski/">Aleksander Hołyński</a>,
<a href="https://roxanneluo.github.io/">Xuan Luo</a>, and Haley Cho for their support and
help with collecting data. Thanks to <a href="https://zhengqili.github.io/">Zhengqi Li</a> and
<a href="http://www.oliverwang.info/">Oliver Wang</a> for their help with the NSFF experiments.</p>
</div>
</section> -->
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link" href="https://github.com/robustsam/RobustSAM">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/robustsam/RobustSAM" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
</div>
</footer>
<script type="text/javascript" src="./static/slick/slick.min.js"></script>
</body>
</html>