-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
372 lines (336 loc) · 21.9 KB
/
index.html
File metadata and controls
372 lines (336 loc) · 21.9 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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>LiteReality: Graphics-Ready 3D Scene Reconstruction from RGB-D Scans</title>
<!-- Meta tags for SEO and social sharing -->
<meta name="description" content="LiteReality: An automatic pipeline that converts RGB-D scans of indoor environments into graphics-ready scenes with high-quality meshes and PBR materials.">
<meta name="keywords" content="3D reconstruction, RGB-D, computer vision, graphics, virtual reality, indoor scanning">
<meta property="og:title" content="LiteReality: Graphics-Ready 3D Scene Reconstruction from RGB-D Scans">
<meta property="og:description" content="An automatic pipeline that converts RGB-D scans of indoor environments into graphics-ready scenes.">
<meta property="og:type" content="website">
<meta property="og:url" content="https://yourusername.github.io/literality">
<meta property="og:image" content="logo.png">
<link rel="stylesheet" href="style.css">
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap" rel="stylesheet">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0/css/all.min.css">
<!-- Three.js CDN -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/three.js/r150/three.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/three@0.150.1/examples/js/loaders/GLTFLoader.js"></script>
<script src="https://cdn.jsdelivr.net/npm/three@0.150.1/examples/js/controls/PointerLockControls.js"></script>
</head>
<body>
<!-- Navigation -->
<nav class="navbar">
<div class="nav-container">
<div class="nav-brand">
<a href="#home">LiteReality</a>
</div>
<ul class="nav-menu">
<li><a href="#home">Home</a></li>
<li><a href="#gallery">Gallery</a></li>
<li><a href="#video">Video</a></li>
<li><a href="#examples">Examples</a></li>
<li><a href="#pipeline">Pipeline</a></li>
<li><a href="#applications">Applications</a></li>
<li><a href="#repainting">PBR Painting</a></li>
<li><a href="#limitations">Limitations</a></li>
<li><a href="#related-work">Related Work</a></li>
<li><a href="#acknowledgments">Acknowledgments</a></li>
<li><a href="#citation">Citation</a></li>
</ul>
</div>
</nav>
<!-- Hero Section -->
<section id="home" class="hero">
<div class="container">
<div class="hero-content">
<div class="hero-title-row">
<img src="logo.png" alt="LiteReality Logo" class="hero-logo">
<span class="hero-title-main">LiteReality</span>
</div>
<h1 class="hero-title-sub">Graphics-Ready 3D Scene Reconstruction from RGB-D Scans</h1>
<div class="hero-links">
<a href="https://arxiv.org/abs/2507.02861" class="btn btn-secondary" target="_blank"><i class="fas fa-file-alt"></i> Paper</a>
<a href="#video" class="btn btn-secondary"><i class="fas fa-play"></i> Video</a>
<a href="https://github.com/LiteReality/LiteReality" class="btn btn-secondary" target="_blank" title="Code"><i class="fab fa-github"></i> Code</a>
<a href="#citation" class="btn btn-secondary" title="BibTeX"><i class="fas fa-book"></i> BibTeX</a>
</div>
<div class="authors-row">
<div class="authors">
<a href="https://zheninghuang.github.io/" target="_blank">Zhening Huang</a><sup>1</sup>,
<a href="https://xywu.me/" target="_blank">Xiaoyang Wu</a><sup>2</sup>,
<a href="https://www.cl.cam.ac.uk/~fz261/" target="_blank">Fangcheng Zhong</a><sup>1</sup>,
<a href="https://hszhao.github.io/" target="_blank">Hengshuang Zhao</a><sup>2</sup>,
<a href="https://niessnerlab.org/members/matthias_niessner/profile.html" target="_blank">Matthias Nießner</a><sup>3</sup>,
<a href="https://www.eng.cam.ac.uk/profiles/jl221" target="_blank">Joan Lasenby</a><sup>1</sup>
</div>
</div>
<div class="institutions">
<sup>1</sup>University of Cambridge, <sup>2</sup>The University of Hong Kong, <sup>3</sup>Technical University of Munich
</div>
<!-- NeurIPS 2025 -->
<div style="margin-top: 12px; text-align: center;">
<span style="
display: inline-block;
background: #111827;
color: #f9fafb;
padding: 6px 14px;
border-radius: 999px;
font-size: 0.9rem;
font-weight: 600;
letter-spacing: 0.03em;
text-transform: uppercase;
">
NeurIPS 2025
</span>
</div>
<div class="hero-description">
<p>
<p>
We are excited to present <b>LiteReality</b> ✨, an automatic pipeline that converts RGB-D scans of indoor environments into <b>graphics-ready</b> 🏠 scenes. In these scenes, all objects are represented as high-quality meshes with PBR materials 🎨 that match their real-world appearance. The scenes also include articulated objects 🔧 and are ready to integrate into graphics pipelines for rendering 💡 and physics-based interactions 🕹️.
</p>
</p>
</div>
<div class="hero-video">
<div class="video-wrapper">
<video controls autoplay muted loop>
<source src="demo.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
</div>
</div>
<br>
<br>
<br>
</div>
</div>
</section>
<!-- Gallery Section -->
<section id="gallery" class="gallery">
<div class="container">
<h2>Gallery</h2>
<div class="hero-description">
<p>
With high-quality meshes and PBR materials 🎨, the reconstructed scenes integrate seamlessly into rendering pipelines 🖥️. Below are examples of rendered scenes, reconstructed from either real-life scans (📷 left) or public datasets (📚 right).
</p>
</div>
<div class="gallery-grid">
<div class="gallery-item">
<img src="Slide1.png" alt="LiteReality Slide 1" class="gallery-img" data-full="Slide1.png">
</div>
<div class="gallery-item">
<img src="Slide2.png" alt="LiteReality Slide 2" class="gallery-img" data-full="Slide2.png">
</div>
</div>
</div>
</section>
<!-- Video Section -->
<section id="video" class="video-section">
<div class="container">
<h2> Video</h2>
<div class="hero-description">
<p>
🎬 Watch our explainer video now and understand everything in just 4 minutes! ⏱️💡✨
</p>
</div>
<!-- Replace local video with YouTube embed -->
<div class="video-wrapper" style="max-width:900px; margin:0 auto;">
<iframe width="100%" height="500" src="https://www.youtube.com/embed/ecK9m3LXg2c" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen></iframe>
</div>
</div>
</section>
<!-- Abstract Section -->
<section id="abstract" class="abstract">
<div class="container">
<h2>Abstract</h2>
<div class="abstract-text">
We propose LiteReality, a novel pipeline that converts RGB-D scans of indoor environments into compact, realistic, and interactive 3D virtual replicas. LiteReality not only reconstructs scenes that visually resemble reality but also supports key features essential for graphics pipelines—such as object individuality, articulation, high-quality physically based rendering materials, and physically based interaction. At its core, LiteReality first performs scene understanding and parses the results into a coherent 3D layout and objects, with the help of a structured scene graph. It then reconstructs the scene by retrieving the most visually similar 3D artist-crafted models from a curated asset database. Later, the Material Painting module enhances the realism of retrieved objects by recovering high-quality, spatially varying materials. Finally, the reconstructed scene is integrated into a simulation engine with basic physical properties applied to enable interactive behavior. The resulting scenes are compact, editable, and fully compatible with standard graphics pipelines, making them suitable for applications in AR/VR, gaming, robotics, and digital twins. In addition, LiteReality introduces a training-free object retrieval module that achieves state-of-the-art similarity performance, as benchmarked on the Scan2CAD dataset, along with a robust Material Painting module capable of transferring appearances from images of any style to 3D assets—even in the presence of severe misalignment, occlusion, and poor lighting. We demonstrate the effectiveness of LiteReality on both real-life scans and public datasets.
</div>
</div>
</section>
<!-- Examples Section -->
<section id="examples" class="examples">
<div class="container">
<h2>Side-by-Side Comparison</h2>
<div class="hero-description" style="text-align: center;">
<p>
🏠 <strong>Left:</strong> Rendered 3D scene reconstruction<br/>
📸 <strong>Right:</strong> Original RGB capture<br/>
Swipe through the carousel above to see more examples. 🔄
</p>
</div>
<div id="examples-thumbnails" class="applications-thumbnails" style="display: flex; justify-content: center; gap: 18px; margin-bottom: 20px;"></div>
<div class="applications-video-container">
<video id="examples-video-player" controls style="width: 100%; max-width: 900px; max-height: 500px; display: block; margin: 0 auto; background: #000; border-radius: 12px; box-shadow: 0 2px 12px rgba(0,0,0,0.10);">
<source src="" type="video/mp4">
Your browser does not support the video tag.
</video>
<div id="examples-caption" class="application-caption" style="margin-top: 10px; text-align: center; color: #2563eb; font-weight: 500;"></div>
</div>
</div>
</section>
<!-- Pipeline Section -->
<section id="pipeline" class="pipeline">
<div class="container">
<h2>Pipeline of LiteReality</h2>
<div class="pipeline-embed">
<img src="pipeline_main.png " alt="LiteReality Pipeline" class="pipeline-img">
</div>
<p class="pipeline-caption">
With input RGB-D scans, the process begins with scene perception and parsing, where room layouts and 3D object bounding box are detected and parsed into a structured and physically plausible arrangement with the help of a scene graph. In the object reconstruction stage, identical clustering first finds identical objects, and a hierarchical retrieval approach is conducted to match 3D models retrieved from the LiteReality database. The material painting stage retrieves and optimizes PBR materials by referencing observed images. Finally, procedural reconstruction assembles all the information into a graphics-ready environment that features realism and interactivity.
</p>
</div>
</section>
<!-- Object Repainting Section -->
<section id="repainting" class="repainting">
<div class="container">
<h2>Material Painting</h2>
<div class="hero-description">
<p>
One of the <strong>key challenges</strong> in the pipeline—which prior work often struggles with—is
<b>reliably recovering PBR materials at scale</b>. While single-image methods perform well on clear views,
they frequently <strong>fail</strong> on room-level scans under conditions of occlusion, poor lighting,
and geometric misalignment between retrieved objects and input images. To address these limitations, we introduce an <strong>MLLM-based retrieval & optimization framework</strong>
for robust, scalable material recovery.
</p>
<p>👇 <strong>Examples:</strong></p>
<ul>
<li>🖌️📸 <strong>Left:</strong> Model painted using <em>multiple reference images</em></li>
<li>🎨✨ <strong>Right:</strong> Single reference image used to paint <em>three different models</em></li>
<br>
</div>
<div class="repainting-grid">
<div class="repainting-item">
<a href="#" class="lightbox-trigger" data-full="MPSlide1.png">
<img src="MPSlide1.png" alt="Object Repainting 1" class="repainting-img">
</a>
</div>
<div class="repainting-item">
<a href="#" class="lightbox-trigger" data-full="MPSlide2.png">
<img src="MPSlide2.png" alt="Object Repainting 2" class="repainting-img">
</a>
</div>
</div>
</div>
</section>
<!-- Applications Section -->
<section id="applications" class="applications">
<div class="container">
<h2>Applications</h2>
<div class="hero-description">
<p>
Creating interactive, graphics-ready scenes unlocks a wide range of application scenarios.
Here, we showcase a few examples, including flexible relighting, physics-based interactions, and object-level manipulation. These capabilities open the door to even more applications, such as VR/AR experiences, robotics simulation, interior design, and digital twin systems.
</p>
</div>
<div id="applications-thumbnails" class="applications-thumbnails" style="display: flex; justify-content: center; gap: 18px; margin-bottom: 20px;"></div>
<div class="applications-video-container">
<video id="applications-video-player" controls style="width: 100%; max-width: 900px; max-height: 500px; display: block; margin: 0 auto; background: #000; border-radius: 12px; box-shadow: 0 2px 12px rgba(0,0,0,0.10);">
<source src="" type="video/mp4">
Your browser does not support the video tag.
</video>
<div id="applications-caption" class="application-caption" style="margin-top: 10px; text-align: center; color: #2563eb; font-weight: 500;"></div>
</div>
</div>
</section>
<!-- Limitations Section -->
<section id="limitations" class="limitations">
<div class="container">
<h2 style="text-align: center;">Limitations and Future Improvements</h2>
<div class="limitations-content">
<p>
At present, LiteReality aims for typical indoor environments without extreme design variations. To generalize to more diverse scenes, several key limitations need to be addressed:
</p>
<ul>
<li>🔄 <strong>Generalization to complex object relationships:</strong><br>
Object relationships are currently inferred from bounding boxes in the detection stage, and scene parsing is performed agnostically to ensure physically plausible layouts. This prevents the system from handling more intricate configurations—e.g., a sink embedded in a cabinet, which is common in kitchen environments.
</li>
<li>💡 <strong>Lack of lighting estimation:</strong><br>
Our material-painting stage skips any explicit lighting estimation. While this simplifies the pipeline, it also reduces the photorealism of the rendered scenes compared to the original captures.
</li>
<li>🐁 <strong>Exclusion of small objects:</strong><br>
LiteReality currently focuses on reconstructing room-defining elements (similar to Apple's RoomPlan); smaller objects are not yet included in the pipeline.
</li>
</ul>
</div>
</div>
</section>
<!-- Learn More Section -->
<section id="learn-more" class="learn-more">
<div class="container">
<h2>More Details?</h2>
<div class="learn-more-text">
For more details on methodologies, evaluation metrics, comparison with baselines, and the LiteReality database, please refer to our paper.<br><br>
<a href="https://arxiv.org/abs/2507.02861" class="btn btn-primary learn-more-btn" target="_blank" style="font-size:1.2rem; padding: 0.8em 2em; margin-top: 1em;">Read the Paper</a>
</div>
</div>
</section>
<!-- Related Work Section -->
<section id="related-work" class="related-work">
<div class="container">
<h2>Related Work</h2>
<div class="hero-description">
<p>
Some prior and concurrent works that are related to our project — from which we have drawn valuable insights — include (but are not limited to):
</p>
</div>
<div class="related-work-list">
<ul>
<li><a href="https://arxiv.org/abs/2307.03244" target="_blank">PSDR‑Room: Single Photo to Scene using Differentiable Rendering</a></li>
<li><a href="https://arxiv.org/abs/2401.06056" target="_blank">MatSynth: A Modern PBR Materials Dataset</a></li>
<li><a href="https://arxiv.org/abs/2212.04819" target="_blank">Phone2Proc: Bringing Robust Robots Into Our Chaotic World</a></li>
<li><a href="https://arxiv.org/abs/2306.09310" target="_blank">Infinite Photorealistic Worlds Using Procedural Generation (Infinigen)</a></li>
<li><a href="https://arxiv.org/abs/2406.11824" target="_blank">Infinigen Indoors: Photorealistic Indoor Scenes using Procedural Generation</a></li>
<li><a href="https://dl.acm.org/doi/10.1145/3641519.3657504" target="_blank">MaPa: Text-driven Photorealistic Material Painting for 3D Shapes</a></li>
<li><a href="https://arxiv.org/abs/2207.00757" target="_blank">PhotoScene: Photorealistic Material and Lighting Transfer for Indoor Scenes</a></li>
<li><a href="https://arxiv.org/abs/2502.12894" target="_blank">CAST: Component-Aligned 3D Scene Reconstruction from a Single RGB Image</a></li>
<li><a href="https://arxiv.org/abs/2410.07408" target="_blank">Automated Creation of Digital Cousins for Robust Policy Learning</a></li>
<li><a href="https://openaccess.thecvf.com/content/CVPR2025/html/Yu_METASCENES_Towards_Automated_Replica_Creation_for_Real-world_3D_Scans_CVPR_2025_paper.html" target="_blank">MetaScenes: Towards Automated Replica Creation for Real-world 3D Scans</a></li>
</ul>
</div>
</div>
</section>
<!-- Acknowledgments Section -->
<section id="acknowledgments" class="acknowledgments">
<div class="container">
<h2>Acknowledgments</h2>
<div class="acknowledgments-text">
<p>
This project was supported by the Girton College, Cambridge Postgraduate Research Fellowship. We gratefully acknowledge the insightful discussions and feedback from Prof Kwang Moo Yi, Erqun Dong, Weiyang Liu, Prof Ayush Tewari, and Shangzhan Zhang throughout the course of the project. We also appreciate the kind support from Prof. Angela Dai for the voiceover contribution, and we thank Jieying Chen for her help with proofreading the paper.
</p>
</div>
</div>
</section>
<!-- Citation Section -->
<section id="citation" class="citation">
<div class="container">
<h2>Citation</h2>
<div class="citation-box">
<pre><code>@misc{huang2025literealitygraphicsready3dscene,
title={LiteReality: Graphics-Ready 3D Scene Reconstruction from RGB-D Scans},
author={Zhening Huang and Xiaoyang Wu and Fangcheng Zhong and Hengshuang Zhao and Matthias Nießner and Joan Lasenby},
year={2025},
eprint={2507.02861},
url={https://arxiv.org/abs/2507.02861}
}
</code></pre>
</div>
</div>
</section>
<!-- Footer -->
<footer class="footer">
<div class="container">
<p>© 2025 LiteReality Project. All rights reserved.</p>
</div>
</footer>
<!-- Lightbox Modal for Gallery -->
<div id="lightbox-modal" class="lightbox-modal">
<span class="lightbox-close" id="lightbox-close">×</span>
<img class="lightbox-content" id="lightbox-img" src="" alt="High Resolution Gallery Image">
</div>
<script src="script.js"></script>
<script src="threejs-viewer.js"></script>
</body>
</html>