LION/index.html
2022-09-21 02:22:48 -04:00

780 lines
26 KiB
HTML
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<script type="text/javascript" charset="utf-8" src="https://ajax.googleapis.com/ajax/libs/jquery/1.3.2/jquery.min.js"></script>
<script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
<style type="text/css">
body {
font-family: "Titillium Web", "HelveticaNeue-Light", "Helvetica Neue Light", "Helvetica Neue", Helvetica, Arial, "Lucida Grande", sans-serif;
font-weight: 300;
font-size: 17px;
margin-left: auto;
margin-right: auto;
}
@media screen and (min-width: 980px){
body {
width: 980px;
}
}
h1 {
font-weight:300;
line-height: 1.15em;
}
h2 {
font-size: 1.75em;
}
a:link,a:visited {
color: #5364cc;
text-decoration: none;
}
a:hover {
color: #208799;
}
h1 {
text-align: center;
}
h2,h3 {
text-align: left;
}
h1 {
font-size: 40px;
font-weight: 500;
}
h2 {
font-weight: 400;
margin: 16px 0px 4px 0px;
}
h3 {
font-weight: 600;
margin: 16px 0px 4px 0px;
}
.paper-title {
padding: 1px 0px 1px 0px;
}
section {
margin: 32px 0px 32px 0px;
text-align: justify;
clear: both;
}
.col-5 {
width: 20%;
float: left;
}
.col-4 {
width: 25%;
float: left;
}
.col-3 {
width: 33%;
float: left;
}
.col-2 {
width: 50%;
float: left;
}
.col-1 {
width: 100%;
float: left;
}
.author-row, .affil-row {
font-size: 26px;
}
.author-row-new {
text-align: center;
}
.author-row-new a {
display: inline-block;
font-size: 20px;
padding: 4px;
}
.author-row-new sup {
color: #313436;
font-size: 12px;
}
.affiliations-new {
font-size: 18px;
text-align: center;
width: 80%;
margin: 0 auto;
margin-bottom: 20px;
}
.row {
margin: 16px 0px 16px 0px;
}
.authors {
font-size: 26px;
}
.affiliatons {
font-size: 18px;
}
.affil-row {
margin-top: 18px;
}
.teaser {
max-width: 100%;
}
.text-center {
text-align: center;
}
.screenshot {
width: 256px;
border: 1px solid #ddd;
}
.screenshot-el {
margin-bottom: 16px;
}
hr {
height: 1px;
border: 0;
border-top: 1px solid #ddd;
margin: 0;
}
.material-icons {
vertical-align: -6px;
}
p {
line-height: 1.25em;
}
.caption {
font-size: 16px;
color: #666;
margin-top: 4px;
margin-bottom: 10px;
}
video {
display: block;
margin: auto;
}
figure {
display: block;
margin: auto;
margin-top: 10px;
margin-bottom: 10px;
}
#bibtex pre {
font-size: 14px;
background-color: #eee;
padding: 16px;
}
.blue {
color: #2c82c9;
font-weight: bold;
}
.orange {
color: #d35400;
font-weight: bold;
}
.flex-row {
display: flex;
flex-flow: row wrap;
padding: 0;
margin: 0;
list-style: none;
}
.paper-btn-coming-soon {
position: relative;
top: 0;
left: 0;
}
.coming-soon {
position: absolute;
top: -15px;
right: -15px;
}
.paper-btn {
position: relative;
text-align: center;
display: inline-block;
margin: 8px;
padding: 8px 8px;
border-width: 0;
outline: none;
border-radius: 2px;
background-color: #5364cc;
color: white !important;
font-size: 20px;
width: 100px;
font-weight: 600;
}
.paper-btn-parent {
display: flex;
justify-content: center;
margin: 16px 0px;
}
.paper-btn:hover {
opacity: 0.85;
}
.container {
margin-left: auto;
margin-right: auto;
padding-left: 16px;
padding-right: 16px;
}
.venue {
font-size: 23px;
}
.topnav {
background-color: #EEEEEE;
overflow: hidden;
}
.topnav div {
max-width: 1070px;
margin: 0 auto;
}
.topnav a {
display: inline-block;
color: black;
text-align: center;
vertical-align: middle;
padding: 16px 16px;
text-decoration: none;
font-size: 18px;
}
.topnav img {
padding: 2px 0px;
width: 100%;
margin: 0.2em 0px 0.3em 0px;
vertical-align: middle;
}
pre {
font-size: 0.9em;
padding-left: 7px;
padding-right: 7px;
padding-top: 3px;
padding-bottom: 3px;
border-radius: 3px;
background-color: rgb(235, 235, 235);
overflow-x: auto;
}
.download-thumb {
display: flex;
}
@media only screen and (max-width: 620px) {
.download-thumb {
display: none;
}
}
.paper-stuff {
width: 50%;
font-size: 20px;
}
@media only screen and (max-width: 620px) {
.paper-stuff {
width: 100%;
}
}
* {
box-sizing: border-box;
}
.column {
text-align: center;
float: left;
width: 16.666%;
padding: 5px;
}
/* Clearfix (clear floats) */
.row::after {
content: "";
clear: both;
display: table;
}
/* Responsive layout - makes the three columns stack on top of each other instead of next to each other */
@media screen and (max-width: 500px) {
.column {
width: 100%;
}
}
</style>
<script type="text/javascript" src="../js/hidebib.js"></script>
<link href='https://fonts.googleapis.com/css?family=Titillium+Web:400,600,400italic,600italic,300,300italic' rel='stylesheet' type='text/css'>
<head>
<title> LION: Latent Point Diffusion Models for 3D Shape Generation </title>
<meta name="viewport" content="width=device-width, initial-scale=1">
<!-- meta property="og:description" content="Score-Based Generative Modeling with Critically-Damped Langevin Diffusion"/ -->
<link href="https://fonts.googleapis.com/css2?family=Material+Icons" rel="stylesheet">
<!-- meta name="twitter:card" content="summary_large_image" -->
<!-- meta name="twitter:creator" content="@timudk" -->
<!-- meta name="twitter:title" content="Score-Based Generative Modeling with Critically-Damped Langevin Diffusion" -->
<!-- meta name="twitter:description" content="Inspired by connections to statistical mechanics, we propose a novel diffusion process, critically-damped Langevin diffusion, that perturbs the data in a smoother manner by leveraging auxiliary velocity variables. This allows us to denoise more efficiently and learn higher quality generative models." -->
<!-- meta name="twitter:image" content="https://nv-tlabs.github.io/CLD-SGM/assets/cld_teaser_resized.png" -->
</head>
<body>
<div class="topnav" id="myTopnav">
<div>
<a href="https://www.nvidia.com/"><img width="100%" src="assets/nvidia.svg"></a>
<a href="https://nv-tlabs.github.io/" ><strong>Toronto AI Lab</strong></a>
</div>
</div>
<div class="container">
<div class="paper-title">
<h1>
<font color="#5364cc">LION</font>:
<font color="#5364cc">L</font>atent Point Diffus<font color="#5364cc">ion</font> Models <br> for 3D Shape Generation</h1>
</div>
<div id="authors">
<center>
<div class="author-row-new">
<a href="https://www.cs.utoronto.ca/~xiaohui/">Xiaohui Zeng<sup>1,2,3</sup></a>,
<a href="http://latentspace.cc/">Arash Vahdat<sup>1</sup></a>,
<a href="https://www.fwilliams.info/">Francis Williams<sup>1</sup></a>,
<a href="https://zgojcic.github.io/">Zan Gojcic<sup>1</sup></a>,
<a href="https://orlitany.github.io/">Or Litany<sup>1</sup></a>,
<a href="https://www.cs.utoronto.ca/~fidler/">Sanja Fidler<sup>1,2,3</sup></a>,
<a href="https://karstenkreis.github.io/">Karsten Kreis<sup>1</sup></a>
</div>
</center>
<center>
<div class="affiliations">
<span><sup>1</sup> NVIDIA</span>
<span><sup>2</sup> University of Toronto</span>
<span><sup>3</sup> Vector Institute</span> <br/>
</div>
<div class="affil-row">
<div class="venue text-center"><b>NeurIPS 2022 </b></div>
</div>
</center>
<div style="clear: both">
<div class="paper-btn-parent">
<a class="paper-btn" href="https://arxiv.org/abs/2112.07068">
<span class="material-icons"> description </span>
Paper
</a>
<div class="paper-btn-coming-soon">
<a class="paper-btn" href="https://github.com/nv-tlabs/LION">
<span class="material-icons"> code </span>
Code
</a>
</div>
</div></div>
</div>
<br>
<section id="teaser-image">
<center>
</p><figure style="margin-top: 20px; margin-bottom: 20px;">
<img width="50%" src="./assets/lion_teaser.png" style="margin-bottom: 20px;">
<p class="caption">
caption for the teaser
</p><p class="caption">
</p>
</center>
</section>
<section id="news">
<hr>
<h2>News</h2>
<div class="row">
<div><span class="material-icons"> event </span> [Sept 2022] Build the project page <a href="https://github.com/nv-tlabs/LION">Page</a>!</div>
</div>
</section>
<section id="abstract"/>
<hr>
<h2>Abstract</h2>
<div class="flex-row">
<p>
Denoising diffusion models (DDMs) have shown promising results in 3D point cloud synthesis. To advance 3D DDMs and make them useful
for digital artists, we require (i) high generation quality, (ii) flexibility for manipulation and applications such as conditional
synthesis and shape interpolation, and (iii) the ability to output smooth surfaces or meshes. To this end, we introduce the
hierarchical Latent Point Diffusion Model (LION) for 3D shape generation. LION is set up as a variational autoencoder (VAE) with
a hierarchical latent space that combines a global shape latent representation with a point-structured latent space. For generation,
we train two hierarchical DDMs in these latent spaces. The hierarchical VAE approach boosts performance compared to DDMs that operate
on point clouds directly, while the point-structured latents are still ideally suited for DDM-based modeling. Experimentally, LION
achieves state-of-the-art generation performance on multiple ShapeNet benchmarks. Furthermore, our VAE framework allows us to easily
use LION for different relevant tasks without re-training the latent DDMs: We show that LION excels at multimodal shape denoising and
voxel-conditioned synthesis. We also demonstrate shape autoencoding and latent shape interpolation, and we augment LION with modern
surface reconstruction techniques to generate smooth 3D meshes. We hope that LION provides a powerful tool for artists working with
3D shapes due to its high-quality generation, flexibility, and surface reconstruction.
</p>
</div>
</section>
<section id="method"/>
<hr>
<h2>Method</h2>
<div class="flex-row">
<p>
LION is set up as a hierarchical point cloud VAE with denoising diffusion models over the shape latent and latent point distributions.
Point-Voxel CNNs (PVCNN) with adaptive Group Normalization (Ada. GN) are used as neural networks.
The latent points can be interpreted as a smoothed version of the input point cloud.
Shape As Points (SAP) is optionally used for mesh reconstruction.
</p>
</div>
<center>
<figure style="width: 100%;">
<a>
<img width="50%" src="assets/pipeline.jpg">
</a>
<p class="caption" style="margin-bottom: 24px;">
Architecture of LION.
</p>
</figure>
</center>
</section>
<!--
<section id="teaser-video">
</p>
<center>
<figure>
<video class="centered" width="50%" controls muted autoplay>
<source src="assets/LION_demo.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated output from LION.
</p>
</figure>
</center>
</p>
</section>
-->
<section id="novelties"/>
<hr>
<h2>Technical Contributions</h2>
<div class="flex-row">
<p>We make the following technical contributions:
<ul style="list-style-type:disc;">
<li>We explore the training of multiple denoising diffusion models (DDMs) in a latent space..</li>
<li>We train latent DDMs in 3D generation.</li>
<li>We outperform all baselines and demonstrate that LION scale to extremely diverse shape datasets, like modeling 13 or even 55 ShapeNet categories jointly without conditioning. </li>
</ul>
</p>
</div>
</section>
<section id="results">
<hr>
<h2>Generation (Single Category)</h2>
<div class="flex-row">
<p>Samples from LION trained on single catgory. </p>
</div>
<center>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_airplane.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh of airplanes.
</p> <br>
</figure>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_chair.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh of chair.
</p> <br>
</figure>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_car.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh of car.
</p> <br>
</figure>
<figure style="width: 100%;">
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_animal553_v2.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption" style="margin-bottom: 24px;">
Generated point clouds and reconstructed mesh of Animal.
</p> <br>
</figure>
</center>
<center>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_bottle.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh of bottle.
</p> <br>
</figure>
</center>
<center>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_mug.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh of mug.
</p> <br>
</figure>
</center>
<hr>
<h2>Generation (Multi-Classes)</h2>
<!-- <div class="flex-row">
<p>samples from LION trained on multiple ShapeNet catgories, without conditioning. </p>
</div> -->
<center>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_all_v13.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh. LION model trained on 13 ShapeNet categories jointly without conditioning.
</p>
<br>
</figure>
</center>
<center>
<figure>
<video class="centered" width="100%" controls autoplay muted playsinline class="video-background " >
<source src="assets/gen_all_55.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Generated point clouds and reconstructed mesh. LION model trained on 55 ShapeNet categories jointly without conditioning.
</p>
<br>
</figure>
</center>
</section>
<section id="more_results">
<hr>
<h2>Applications</h2>
<h3>Interpolation </h3>
<div class="flex-row">
<p>LION can interpolate two shapes by traversing the latent space. The generated shapes are clean and semantically plausible along the entire interpolation path. </p>
</div>
<figure>
<video class="centered" width="100%" controls muted playsinline class="video-background " >
<source src="assets/LION_interp.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
Left most shape: the source shape. Right most shape: the target shape. The shapes in middle are interpolated results between source and target shape.
</p>
</figure>
<center>
<figure>
<video class="centered" width="50%" controls loop autoplay muted playsinline class="video-background " >
<source src="assets/LION_interp_seq.mp4#t=0.001" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption">
LION traverses the latent space and interpolates many different shapes.
</p>
</figure>
</center>
<br>
<h3>Voxel-Conditioned Synthesis </h3>
<div class="flex-row">
<p>Given a coarse voxel grid, LION can generate different plausible detailed shapes. </p>
<p>In practice, an artist using a 3D generative model may have a rough idea of the desired shape. For instance, they may be able to quickly construct a coarse voxelized shape, to which the generative model then adds realistic details. </p>
</div>
<center>
<figure style="width: 80%;">
<video class="centered" width="80%" controls muted playsinline class="video-background " >
<source src="assets/airplane_voxel.mp4#t=0.14" type="video/mp4">
Your browser does not support the video tag.
</video>
<p class="caption" style="margin-bottom: 24px;" width="30%">
Left: Input voxel grid. Right: two point clouds generated by LION and the reconstructed mesh.
<!-- Voxel-guided synthesis experiments, on different categories. We run diffuse-denoise in latent space to generate diverse plausible clean shapes (first row, left plane: 250 diffuse-denoise steps; first row, right plane: 200 steps;) -->
</p>
</figure>
</center>
<br>
<h3> Single View Reconstruction </h3>
<div class="flex-row">
<p>
We extend LION to also allow for single view reconstruction (SVR) from RGB data. We render 2D
images from the 3D ShapeNet shapes, extracted the images CLIP image embeddings, and
trained LIONs latent diffusion models while conditioning on the shapes CLIP image embeddings.
At test time, we then take a single view 2D image, extract the CLIP image embedding, and generate
corresponding 3D shapes, thereby effectively performing SVR. We show SVR results from real
RGB data
</p>
</div>
<center>
<figure style="width: 100%;">
<a>
<img width="49%" src="assets/svr/img2shape_mitsuba_full.jpg">
<img width="49%" src="assets/svr/img2shape_cari2s_mm_mitsuba_full.jpg">
</a>
<p class="caption" style="margin-bottom: 24px;">
Single view reconstruction from RGB images of chair. For each input image, LION can generate multi-modal outputs.
</p>
</figure>
</center>
<!--
<figure style="width: 50%;">
<a>
<img width="100%" src="assets/svr/img2shape_cari2s_mm_mitsuba_full.jpg">
</a>
<p class="caption" style="margin-bottom: 24px;">
Single view reconstruction from RGB images of car. For each input image, LION can generate multi-modal outputs.
</p>
</figure>
-->
<center>
<figure style="width: 100%;">
<a>
<img width="100%" src="assets/svr/img2shape_cari2s_mitsuba_full.jpg">
</a>
<p class="caption" style="margin-bottom: 24px;">
More single view reconstruction from RGB images of car.
</p>
</figure>
</center>
<br>
<h3> Text-Guided Generation </h3>
<div class="flex-row">
<p>
Using CLIPs text encoder, our method additionally allows for text-guided generation.
</p>
</div>
<center>
<figure style="width: 100%;">
<a>
<img width="35%" src="assets/clipforge_chair.png">
<img width="60%" src="assets/clipforge_car.png">
</a>
<p class="caption" style="margin-bottom: 24px;">
Text-driven shape generation of chairs with LION. Bottom row is the text input
</p>
</figure>
</center>
<h3> Per-sample Text-driven Texture Synthesis </h3>
<div class="flex-row">
<p>
We apply Text2mesh on some generated meshes from LION to additionally synthesize textures in a text-driven manner, leveraging CLIP. The original mesh is generated by LION.
</p>
</div>
<div class="row">
<div class="column">
<img width="100%" src="assets/text2mesh/strawberries_airplane-rec_3.jpg">
<figcaption align = "center">An airplane made of strawberry</figcaption>
</div>
<div class="column">
<img width="100%" src="assets/text2mesh/fabric_leather_airplane-rec_3.jpg">
<figcaption align = "center">An airplane made of fabric leather </figcaption>
</div>
<div class="column">
<img width="100%" src="assets/text2mesh/wood_chair-rec_421_norm1.jpg">
<figcaption align = "center">A chair made of wood</figcaption>
</div>
<div class="column">
<img width="100%" src="assets/text2mesh/wrong_copied1-rec_293_norm0.jpg">
<figcaption align = "center">A car made of rusty metal</figcaption>
</div>
<div class="column">
<img width="100%" src="assets/text2mesh/brick_car-rec_67_norm1.jpg">
<figcaption align = "center">A car made of brick</figcaption>
</div>
<div class="column">
<img width="100%" src="assets/text2mesh/wrong_copied1-rec_12_norm1.jpg">
<figcaption align = "center">A denim fabric animal</figcaption>
</div>
</div>
<br>
</section>
<section id="paper">
<h2>Paper</h2>
<hr>
<div class="flex-row">
<div class="download-thumb">
<div style="box-sizing: border-box; padding: 16px; margin: auto;">
<a href="https://nv-tlabs.github.io/CLD-SGM"><img class="screenshot" src="assets/cld_paper_preview.png"></a>
</div>
</div>
<div class="paper-stuff">
<p><b>LION: Latent Point Diffusion Models for 3D Shape Generation</b></p>
<p>Xiaohui Zeng, Arash Vahdat, Francis Williams, Zan Gojcic, Or Litany, Sanja Fidler, Karsten Kreis</p>
<p><i>Advances in Neural Information Processing Systems (NeurIPS), 2022 <b></b></i></p>
<!--
<div><span class="material-icons"> description </span><a href="https://arxiv.org/abs/2112.07068"> arXiv version</a></div>
<div><span class="material-icons"> insert_comment </span><a href="assets/dockhorn2021score.bib"> BibTeX</a></div>
<div><span class="material-icons"> integration_instructions </span><a href="https://github.com/nv-tlabs/CLD-SGM"> Code</a></div>
-->
</div>
</div>
</div>
</section>
<section id="bibtex">
<h2>Citation</h2>
<hr>
<pre><code>@inproceedings{
zeng2022lion,
title={ LION: Latent Point Diffusion Models for 3D Shape Generation },
author={ Xiaohui Zeng and Arash Vahdat and Francis Williams and Zan Gojcic and Or Litany and Sanja Fidler and Karsten Kreis },
booktitle={Advances in Neural Information Processing Systems (NeurIPS)},
year={2022}
}</code></pre>
</section>
</div>
</body>
</html>