-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
522 lines (463 loc) · 18.1 KB
/
Copy pathindex.html
File metadata and controls
522 lines (463 loc) · 18.1 KB
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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="AquaDiff: A diffusion-based underwater image enhancement framework with chromatic prior guidance and cross-domain consistency.">
<meta name="keywords"
content="Underwater Image Enhancement, Diffusion Model, Chromatic Prior, Cross-Attention, Multi-Resolution Attention, Cross-Domain Consistency">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>AquaDiff: Diffusion-Based Underwater Image Enhancement</title>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" rel="stylesheet">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/bulma/0.9.4/css/bulma.min.css">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.1/css/all.min.css">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<style>
/* PaperBanana-inspired styling */
:root {
--primary-color: #1a3b5d;
--accent-color: #3a7ca5;
--light-bg: #f8f9fa;
--border-radius: 12px;
--box-shadow: 0 10px 30px rgba(0,0,0,0.1);
--ocean-gradient: linear-gradient(135deg, #0a2f44 0%, #1b5e7a 100%);
}
body {
font-family: 'Google Sans', 'Noto Sans', sans-serif;
line-height: 1.6;
color: #2c3e50;
}
.hero {
background: var(--ocean-gradient);
color: white;
padding: 4rem 1.5rem;
}
.hero .title {
color: white !important;
font-size: 2.8rem !important;
font-weight: 400 !important;
margin-bottom: 1.5rem !important;
}
.hero .subtitle {
color: rgba(255,255,255,0.9) !important;
font-size: 1.2rem !important;
}
.section {
padding: 4rem 1.5rem;
}
.section-title {
font-size: 2.2rem;
font-weight: 400;
color: var(--primary-color);
margin-bottom: 2rem;
text-align: center;
}
.content-card {
background: white;
border-radius: var(--border-radius);
padding: 2rem;
box-shadow: var(--box-shadow);
margin-bottom: 2rem;
}
.publication-authors {
font-size: 1.2rem;
margin: 1.5rem 0;
}
.author-block {
margin: 0 0.5rem;
}
.publication-links {
margin: 2rem 0;
}
.button.is-rounded {
border-radius: 30px;
padding: 0.5rem 1.8rem;
font-size: 1rem;
transition: all 0.3s ease;
background-color: rgba(255,255,255,0.2);
color: white;
border: 1px solid rgba(255,255,255,0.3);
}
.button.is-rounded:hover {
transform: translateY(-2px);
box-shadow: 0 5px 15px rgba(0,0,0,0.2);
background-color: rgba(255,255,255,0.3);
color: white;
}
.teaser-image {
border-radius: var(--border-radius);
box-shadow: var(--box-shadow);
max-width: 100%;
height: auto;
margin: 1rem 0;
}
.results-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(350px, 1fr));
gap: 2rem;
margin: 2rem 0;
}
.results-item {
text-align: center;
background: white;
border-radius: var(--border-radius);
padding: 1.5rem;
box-shadow: var(--box-shadow);
transition: transform 0.3s ease;
}
.results-item:hover {
transform: translateY(-5px);
}
.results-item img {
width: 100%;
border-radius: 8px;
margin-bottom: 1rem;
}
.caption {
font-size: 0.95rem;
color: #666;
margin-top: 0.5rem;
line-height: 1.4;
}
.comparison-row {
display: flex;
gap: 2rem;
margin: 2rem 0;
flex-wrap: wrap;
justify-content: center;
}
.comparison-col {
flex: 1;
min-width: 300px;
background: white;
border-radius: var(--border-radius);
padding: 1.5rem;
box-shadow: var(--box-shadow);
}
.comparison-col img {
width: 100%;
border-radius: 8px;
margin-bottom: 1rem;
}
.feature-list {
list-style: none;
padding: 0;
}
.feature-list li {
margin: 1rem 0;
padding-left: 1.5rem;
position: relative;
}
.feature-list li:before {
content: "▹";
color: var(--accent-color);
position: absolute;
left: 0;
}
.abstract-text {
font-size: 1.1rem;
text-align: justify;
max-width: 800px;
margin: 0 auto;
}
.bibtex-code {
background: var(--light-bg);
border-radius: var(--border-radius);
padding: 1.5rem;
font-family: monospace;
font-size: 0.9rem;
overflow-x: auto;
}
.footer {
background: var(--light-bg);
padding: 3rem 1.5rem;
margin-top: 2rem;
}
.footer a {
color: var(--primary-color);
margin: 0 0.5rem;
}
.footer a:hover {
color: var(--accent-color);
}
.institution {
font-size: 1.1rem;
color: rgba(255,255,255,0.8);
margin-top: 1rem;
}
.full-width-image {
width: 100%;
max-width: 1200px;
margin: 0 auto;
display: block;
}
.table-container {
background: white;
border-radius: var(--border-radius);
padding: 2rem;
box-shadow: var(--box-shadow);
margin: 2rem 0;
overflow-x: auto;
}
.table-container img {
width: 100%;
height: auto;
border-radius: 8px;
}
.grid-2col {
display: grid;
grid-template-columns: repeat(2, 1fr);
gap: 2rem;
}
@media screen and (max-width: 768px) {
.hero .title {
font-size: 2rem !important;
}
.section-title {
font-size: 1.8rem;
}
.results-grid {
grid-template-columns: 1fr;
}
.grid-2col {
grid-template-columns: 1fr;
}
}
</style>
</head>
<body>
<!-- HERO SECTION -->
<section class="hero">
<div class="container">
<div class="columns is-centered">
<div class="column is-10 has-text-centered">
<h1 class="title is-1">
<b>AquaDiff</b> 🌊
</h1>
<h2 class="subtitle is-3">
Diffusion-Based Underwater Image Enhancement with Chromatic Prior Guidance and Cross-Domain Consistency
</h2>
<div class="publication-authors">
<span class="author-block">Afrah Shaahid<sup>1,2</sup>,</span>
<span class="author-block">Muzammil Behzad<sup>1,2</sup></span>
</div>
<div class="institution">
<sup>1</sup>King Fahd University of Petroleum and Minerals · <sup>2</sup>SDAIA-KFUPM Joint Research Center for Artificial Intelligence
</div>
<div class="publication-links">
<span class="link-block">
<a href="#" target="_blank" class="button is-rounded is-light">
<span class="icon"><i class="fas fa-file-pdf"></i></span>
<span>Paper</span>
</a>
</span>
<span class="link-block">
<a href="#" target="_blank" class="button is-rounded is-light">
<span class="icon"><i class="ai ai-arxiv"></i></span>
<span>arXiv</span>
</a>
</span>
<span class="link-block">
<a href="https://github.com/BRAIN-Lab-AI/AquaDiff" target="_blank" class="button is-rounded is-light">
<span class="icon"><i class="fab fa-github"></i></span>
<span>Code</span>
</a>
</span>
</div>
</div>
</div>
</div>
</section>
<!-- TEASER / ARCHITECTURE -->
<section class="section">
<div class="container">
<div class="content has-text-centered">
<img src="static/images/AquaDiffArchitecture.png" alt="AquaDiff Architecture" class="teaser-image" style="max-width: 100%;">
<p class="caption">
Overview of the proposed AquaDiff framework. The forward diffusion process progressively corrupts the clean reference image, while the reverse diffusion process with cross-attention conditioning and chromatic prior guidance recovers the enhanced underwater image.
</p>
</div>
</div>
</section>
<!-- ABSTRACT -->
<section class="section" style="background: var(--light-bg);">
<div class="container">
<h2 class="section-title">Abstract</h2>
<div class="content-card">
<div class="abstract-text">
<p>
Underwater images are severely degraded by wavelength-dependent light absorption and scattering, resulting in color distortion, low contrast, and loss of fine details that hinder vision-based underwater applications. To address these challenges, we propose <strong>AquaDiff</strong>, a diffusion-based underwater image enhancement framework designed to correct chromatic distortions while preserving structural and perceptual fidelity. AquaDiff integrates a chromatic prior-guided color compensation strategy with a conditional diffusion process, where cross-attention dynamically fuses degraded inputs and noisy latent states at each denoising step. An enhanced denoising backbone with residual dense blocks and multi-resolution attention captures both global color context and local details. Furthermore, a novel cross-domain consistency loss jointly enforces pixel-level accuracy, perceptual similarity, structural integrity, and frequency-domain fidelity. Extensive experiments on multiple challenging underwater benchmarks demonstrate that AquaDiff achieves superior color correction and competitive overall image quality across diverse underwater conditions.
</p>
</div>
</div>
</div>
</section>
<!-- KEY COMPONENTS -->
<section class="section">
<div class="container">
<h2 class="section-title">Key Components</h2>
<div class="grid-2col">
<div class="content-card">
<h3 class="title is-4" style="color: var(--primary-color);">🔵 Chromatic Prior-Guided Color Compensation</h3>
<p>Preprocesses degraded images using 3-channel compensation (3C) in Lab color space to correct color casts before diffusion conditioning, with spatially-varying masks to prevent overcompensation.</p>
</div>
<div class="content-card">
<h3 class="title is-4" style="color: var(--primary-color);">🟢 Cross-Attention Conditioning</h3>
<p>Dynamically fuses noisy intermediate states with color-compensated conditioning images at each denoising step, enabling adaptive feature weighting based on noise levels.</p>
</div>
<div class="content-card">
<h3 class="title is-4" style="color: var(--primary-color);">🟠 Enhanced Denoising Backbone</h3>
<p>U-Net architecture with residual dense blocks, dense skip connections, and multi-resolution attention modules to capture both global color context and local structural details.</p>
</div>
<div class="content-card">
<h3 class="title is-4" style="color: var(--primary-color);">🟣 Cross-Domain Consistency Loss</h3>
<p>Novel hybrid loss jointly enforcing pixel-level accuracy, perceptual similarity, structural integrity (SSIM), and frequency-domain fidelity (FFT magnitude spectrum matching).</p>
</div>
</div>
</div>
</section>
<!-- COMPARISON RESULTS -->
<section class="section" style="background: var(--light-bg);">
<div class="container">
<h2 class="section-title">Comparison with State-of-the-Art</h2>
<div class="comparison-row">
<div class="comparison-col">
<img src="static/images/U90.png" alt="U90 dataset comparison">
<p class="caption">Results on U90 dataset (TEST-U90) showing superior color correction and detail recovery across diverse underwater scenes.</p>
</div>
<div class="comparison-col">
<img src="static/images/U45.png" alt="U45 dataset comparison">
<p class="caption">Results on U45 dataset demonstrating effective correction of green-yellow and blue-green color casts.</p>
</div>
</div>
<div class="comparison-row">
<div class="comparison-col">
<img src="static/images/C60.png" alt="C60 dataset comparison">
<p class="caption">Results on C60 dataset for severely degraded images where reference images are unavailable.</p>
</div>
<div class="comparison-col">
<img src="static/images/S16.png" alt="S16 dataset comparison">
<p class="caption">Results on S16 dataset with color calibration charts showing precise color restoration.</p>
</div>
</div>
</div>
</section>
<!-- QUANTITATIVE RESULTS -->
<section class="section">
<div class="container">
<h2 class="section-title">Quantitative Results</h2>
<!-- Results Table -->
<div class="table-container">
<img src="static/images/results_table.png" alt="Quantitative results table" class="full-width-image">
<p class="caption has-text-centered">
Quantitative comparison on U45, S16, and C60 datasets using UIQM and UCIQE metrics. AquaDiff achieves the highest UCIQE scores across all datasets, demonstrating superior color correction.
</p>
</div>
<!-- Radar Chart -->
<div class="results-grid">
<div class="results-item">
<img src="static/images/AquaDiffradarbars.png" alt="Radar chart performance">
<p class="caption">Radar chart illustrating AquaDiff's balanced performance across multiple quality metrics, with consistent superiority in color correction (UCIQE).</p>
</div>
<!-- Ablation Study Table -->
<div class="results-item">
<h3 class="title is-5" style="color: var(--primary-color); margin-bottom: 1rem;">Ablation Studies</h3>
<table class="table is-bordered is-fullwidth" style="margin-top: 1rem;">
<thead>
<tr>
<th>Model Variant</th>
<th>UIQM ↑</th>
<th>UCIQE ↑</th>
</tr>
</thead>
<tbody>
<tr>
<td>Baseline Diffusion Model</td>
<td>4.12</td>
<td>0.486</td>
</tr>
<tr>
<td>+ CDCL Only</td>
<td>4.38</td>
<td>0.521</td>
</tr>
<tr>
<td>+ Enhanced U-Net Only</td>
<td>4.45</td>
<td>0.528</td>
</tr>
<tr>
<td><strong>AquaDiff (Full Model)</strong></td>
<td><strong>4.61</strong></td>
<td><strong>0.539</strong></td>
</tr>
</tbody>
</table>
<p class="caption">Ablation studies demonstrating the contribution of Cross-Domain Consistency Loss (CDCL) and Enhanced U-Net backbone.</p>
</div>
</div>
</div>
</section>
<!-- DETAILED ANALYSIS -->
<section class="section" style="background: var(--light-bg);">
<div class="container">
<h2 class="section-title">Detailed Analysis</h2>
<div class="content-card">
<h3 class="title is-4" style="color: var(--primary-color); text-align: center;">Color Correction & Detail Preservation</h3>
<div class="comparison-row">
<div class="comparison-col">
<img src="static/images/U90.png" alt="Detail preservation example">
<p class="caption"><strong>Superior color restoration:</strong> AquaDiff effectively corrects severe color casts ranging from dominant blue-green tints to reddish-brown distortions while preserving fine textures and intricate details.</p>
</div>
<div class="comparison-col">
<img src="static/images/C60.png" alt="Haze reduction example">
<p class="caption"><strong>Haze reduction:</strong> AquaDiff achieves superior haze removal with natural appearance, revealing distant objects and fine structural details that remain obscured in competing methods.</p>
</div>
</div>
<div style="margin-top: 2rem;">
<ul class="feature-list">
<li><strong>Blue-green dominant scenes:</strong> AquaDiff consistently restores accurate and natural colors while effectively removing color casts.</li>
<li><strong>Green-yellow turbid scenes:</strong> Reliably neutralizes green-yellow casts while preserving natural sandy tones and realistic color variations.</li>
<li><strong>Reddish-brown turbid environments:</strong> Consistently removes reddish-brown casts and restores natural colors without oversaturation.</li>
<li><strong>Artifact-free outputs:</strong> Produces clean outputs with natural contrast, sharp edges, and coherent brightness distribution.</li>
</ul>
</div>
</div>
</div>
</section>
<!-- BIBTEX -->
<section class="section">
<div class="container">
<h2 class="section-title">BibTeX</h2>
<div class="content-card">
<pre class="bibtex-code"><code>
@article{shaahid2025aquadiff,
title={AquaDiff: Diffusion-Based Underwater Image Enhancement with Chromatic Prior Guidance and Cross-Domain Consistency},
author={Shaahid, Afrah and Behzad, Muzammil},
journal={arXiv preprint arXiv:XXXX.XXXXX},
year={2025}
}
</code></pre>
</div>
</div>
</section>
<!-- FOOTER -->
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<p>
<a href="#" target="_blank"><i class="fas fa-file-pdf fa-2x"></i></a>
<a href="https://github.com/BRAIN-Lab-AI/AquaDiff" target="_blank" style="margin-left: 20px;"><i class="fab fa-github fa-2x"></i></a>
<a href="#" target="_blank" style="margin-left: 20px;"><i class="fas fa-book fa-2x"></i></a>
</p>
<p style="margin-top: 2rem; font-size: 0.9rem; color: #666;">
AquaDiff: Diffusion-Based Underwater Image Enhancement
</p>
<p style="font-size: 0.8rem; color: #999; margin-top: 1rem;">
King Fahd University of Petroleum and Minerals · SDAIA-KFUPM Joint Research Center for Artificial Intelligence
</p>
</div>
</div>
</footer>
</body>
</html>