1
0
DavidZhang73.github.io/IKEA-ASM-Dataset/index.html

510 lines
21 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>
<html lang="zh-CN" class="loading">
<head>
<!-- hexo-inject:begin --><!-- hexo-inject:end --><meta charset="UTF-8" />
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1" />
<meta name="viewport" content="width=device-width, minimum-scale=1.0, maximum-scale=1.0, user-scalable=no">
<title>宜家家居组装数据集 - Blog</title>
<meta name="apple-mobile-web-app-capable" content="yes" />
<meta name="apple-mobile-web-app-status-bar-style" content="black-translucent">
<meta name="google" content="notranslate" />
<meta name="keywords" content="Developer, Python, C++, JavaScript, Java,">
<meta name="description" content="DavidZ&#39;s Blog,
原创发表于 DavidZ Blog遵循 CC 4.0 BY-NC-SA 版权协议,转载请附上原文出处链接及本声明。
¶简介
Title
The IKEA ASM Dataset: Und,">
<meta name="author" content="DavidZ">
<link rel="alternative" href="atom.xml" title="Blog" type="application/atom+xml">
<link rel="icon" href="/img/favicon.png">
<link rel="stylesheet" href="//at.alicdn.com/t/font_1429596_nzgqgvnmkjb.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/animate.css@3.7.2/animate.min.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/social-share.js@1.0.16/dist/css/share.min.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/codemirror@5.48.4/lib/codemirror.min.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/codemirror@5.48.4/theme/dracula.css">
<link rel="stylesheet" href="/css/obsidian.css">
<link rel="stylesheet" href="/css/ball-atom.min.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/font-awesome/css/font-awesome.min.css">
<meta name="generator" content="Hexo 4.2.1"><!-- hexo-inject:begin --><!-- hexo-inject:end --></head>
<body class="loading">
<!-- hexo-inject:begin --><!-- hexo-inject:end --><div class="loader">
<div class="la-ball-atom la-2x">
<div></div>
<div></div>
<div></div>
<div></div>
</div>
</div>
<span id="config-title" style="display:none">Blog</span>
<div id="loader"></div>
<div id="single">
<div class="scrollbar gradient-bg-rev"></div>
<div id="top" style="display: block;">
<div class="bar" style="width: 0;"></div>
<div class="navigation animated fadeIn fast delay-1s">
<img id="home-icon" class="icon-home" src="/img/favicon.png" alt="" data-url="https://blog.davidz.cn">
<div id="play-icon" title="Play/Pause" class="iconfont icon-play"></div>
<h3 class="subtitle">宜家家居组装数据集</h3>
<div class="social">
<!-- <div class="like-icon">-->
<!-- <a href="javascript:;" class="likeThis active"><span class="icon-like"></span><span class="count">76</span></a>-->
<!-- </div>-->
<div>
<div class="share">
<a href="javascript:;" class="iconfont icon-share1"></a>
<div class="share-component-cc" data-disabled="facebook,douban,linkedin,diandian,tencent,google"></div>
</div>
</div>
</div>
</div>
</div>
<div class="section">
<div class=article-header-wrapper>
<div class="article-header">
<div class="article-cover animated fadeIn" style="
animation-delay: 600ms;
animation-duration: 1.2s;
background-image:
radial-gradient(ellipse closest-side, rgba(0, 0, 0, 0.65), #100e17),
url(https://davidz-blog.oss-cn-beijing.aliyuncs.com/img/library-1602131491.jpg) ">
</div>
<div class="else">
<p class="animated fadeInDown">
<a href="/categories/Paper Review"><b>
</b>PAPER REVIEW<b></b></a>
十月 08, 2020
</p>
<h3 class="post-title animated fadeInDown"><a href="/IKEA-ASM-Dataset/" title="宜家家居组装数据集" class="">宜家家居组装数据集</a>
</h3>
<p class="post-count animated fadeInDown">
<span>
<b class="iconfont icon-text2"></b> <i>文章字数</i>
2.6k
</span>
<span>
<b class="iconfont icon-timer__s"></b> <i>阅读约需</i>
2 mins.
</span>
<span id="busuanzi_container_page_pv">
<b class="iconfont icon-read"></b> <i>阅读次数</i>
<span id="busuanzi_value_page_pv">0</span>
</span>
</p>
<ul class="animated fadeInDown post-tags-list" itemprop="keywords"><li class="animated fadeInDown post-tags-list-item"><a class="animated fadeInDown post-tags-list-link" href="/tags/CV/" rel="tag">CV</a></li><li class="animated fadeInDown post-tags-list-item"><a class="animated fadeInDown post-tags-list-link" href="/tags/Dataset/" rel="tag">Dataset</a></li></ul>
</div>
</div>
</div>
<div class="screen-gradient-after">
<div class="screen-gradient-content">
<div class="screen-gradient-content-inside">
<div class="bold-underline-links screen-gradient-sponsor">
<p>
<span class="animated fadeIn delay-1s"></span>
</p>
</div>
</div>
</div>
</div>
<div class="article">
<div class='main'>
<div class="content markdown animated fadeIn">
<blockquote>
<p>原创发表于 <a href="https://blog.davidz.cn">DavidZ Blog</a>,遵循 <a href="https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode" target="_blank" rel="noopener">CC 4.0 BY-NC-SA</a> 版权协议,转载请附上原文出处链接及本声明。</p>
</blockquote>
<h2 id="简介"><a class="header-anchor" href="#简介"></a>简介</h2>
<table>
<thead>
<tr>
<th style="text-align:center">Title</th>
<th style="text-align:center">The IKEA ASM Dataset: Understanding People Assembling Furniture through Actions, Objects and Pose</th>
</tr>
</thead>
<tbody>
<tr>
<td style="text-align:center">Author</td>
<td style="text-align:center">Yizhak Ben-Shabat</td>
</tr>
<tr>
<td style="text-align:center">Publish Year</td>
<td style="text-align:center">2020</td>
</tr>
<tr>
<td style="text-align:center">Link</td>
<td style="text-align:center"><a href="https://arxiv.org/abs/2007.00394" target="_blank" rel="noopener">https://arxiv.org/abs/2007.00394</a></td>
</tr>
<tr>
<td style="text-align:center">Code</td>
<td style="text-align:center"><a href="https://github.com/IkeaASM/IKEA_ASM_Dataset" target="_blank" rel="noopener">https://github.com/IkeaASM/IKEA_ASM_Dataset</a></td>
</tr>
</tbody>
</table>
<p>这个是我的导师小组里的研究员 Itzik 发布的一个组装宜家家居的视频数据集。其中数据集包括三个视角的 RGBD 图像,原子动作标注,人体姿势标注,物体分割标注,相机标定以及相关的辅助代码。</p>
<h2 id="总结"><a class="header-anchor" href="#总结"></a>总结</h2>
<h3 id="数据集"><a class="header-anchor" href="#数据集"></a>数据集</h3>
<p>背景方面,有 48 个人物5 个场景,场景是动态的,存在有无关人员走过的情况。每个场景下都组装两遍,一次在桌子上,一次在地板上,所以一共是 10 中场景设置。</p>
<p>家具方面4 种家居side table, coffee table, TV bench, and drawer每种三个颜色white, oak, and black。</p>
<p>标注方面,人工标注视频的$1%$左右,人工复核过。</p>
<h3 id="基准"><a class="header-anchor" href="#基准"></a>基准</h3>
<h4 id="Action-recognition"><a class="header-anchor" href="#Action-recognition"></a>Action recognition</h4>
<p><img src="https://davidz-blog.oss-cn-beijing.aliyuncs.com/img/table2-1602211486.png" alt="Action recognition"></p>
<h4 id="Multi-view-and-multi-modal-action-recognition"><a class="header-anchor" href="#Multi-view-and-multi-modal-action-recognition"></a>Multi-view and multi-modal action recognition</h4>
<p><img src="https://davidz-blog.oss-cn-beijing.aliyuncs.com/img/table3-1602211507.png" alt="Multi-view and multi-modal action recognition"></p>
<h4 id="Instance-segmentation"><a class="header-anchor" href="#Instance-segmentation"></a>Instance segmentation</h4>
<h4 id="Multiple-furniture-part-tracking"><a class="header-anchor" href="#Multiple-furniture-part-tracking"></a>Multiple furniture part tracking</h4>
<h4 id="Human-pose"><a class="header-anchor" href="#Human-pose"></a>Human pose</h4>
<!--[if lt IE 9]><script>document.createElement('audio');</script><![endif]-->
<audio id="audio" loop="1" preload="auto" controls="controls"
data-autoplay="false">
<source type="audio/mpeg" src="">
</audio>
<ul id="audio-list" style="display:none">
<li title='0' data-url='/statics/chengdu.mp3'></li>
</ul>
<div class="post-nav">
<hr>
<div class="post-nav-item">上一篇:<a href="/vidat/" rel="prev"
title="在线视频标注工具 - Vidat">在线视频标注工具 - Vidat
</a></div>
<div class="post-nav-item">下一篇:<a href="/acrylic-not-working/" rel="next"
title="技巧速记Windows 10 透明效果无法正常工作">技巧速记Windows 10 透明效果无法正常工作</a></div>
</div>
<div id='gitalk-container' class="comment link"
data-ae='true'
data-ci='489076c5dd3f5ba13f67'
data-cs='d6e3b245787b0b74d0dbe2639ef87f452a401194'
data-r='blog.davidz.cn'
data-o='DavidZhang73'
data-a='DavidZhang73'
data-d=''
data-p='https://cors-anywhere.azm.workers.dev/https://github.com/login/oauth/access_token'
>留言</div>
</div>
<div class="sidebar">
<div class="box animated fadeInRight">
<div class="subbox">
<img src="https://davidz-blog.oss-cn-beijing.aliyuncs.com/img/2019-1599483796.jpg" height=300 width=300></img>
<p>DavidZ</p>
<span>凡事都要留几分</span>
<dl>
</dl>
</div>
<ul>
<li><a href="/">15 <p>文章</p></a></li>
<li><a href="/categories">8 <p>分类</p></a></li>
<li><a href="/tags">15 <p>标签</p></a></li>
</ul>
</div>
<div class="box sticky animated fadeInRight faster">
<div id="toc" class="subbox">
<h4>目录</h4>
<ol class="toc"><li class="toc-item toc-level-2"><a class="toc-link" href="#简介"><span class="toc-number">1.</span> <span class="toc-text">简介</span></a></li><li class="toc-item toc-level-2"><a class="toc-link" href="#总结"><span class="toc-number">2.</span> <span class="toc-text">总结</span></a><ol class="toc-child"><li class="toc-item toc-level-3"><a class="toc-link" href="#数据集"><span class="toc-number">2.1.</span> <span class="toc-text">数据集</span></a></li><li class="toc-item toc-level-3"><a class="toc-link" href="#基准"><span class="toc-number">2.2.</span> <span class="toc-text">基准</span></a><ol class="toc-child"><li class="toc-item toc-level-4"><a class="toc-link" href="#Action-recognition"><span class="toc-number">2.2.1.</span> <span class="toc-text">Action recognition</span></a></li><li class="toc-item toc-level-4"><a class="toc-link" href="#Multi-view-and-multi-modal-action-recognition"><span class="toc-number">2.2.2.</span> <span class="toc-text">Multi-view and multi-modal action recognition</span></a></li><li class="toc-item toc-level-4"><a class="toc-link" href="#Instance-segmentation"><span class="toc-number">2.2.3.</span> <span class="toc-text">Instance segmentation</span></a></li><li class="toc-item toc-level-4"><a class="toc-link" href="#Multiple-furniture-part-tracking"><span class="toc-number">2.2.4.</span> <span class="toc-text">Multiple furniture part tracking</span></a></li><li class="toc-item toc-level-4"><a class="toc-link" href="#Human-pose"><span class="toc-number">2.2.5.</span> <span class="toc-text">Human pose</span></a></li></ol></li></ol></li></ol>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<div id="back-to-top" class="animated fadeIn faster">
<div class="flow"></div>
<span class="percentage animated fadeIn faster">0%</span>
<span class="iconfont icon-top02 animated fadeIn faster"></span>
</div><!-- hexo-inject:begin --><!-- hexo-inject:end -->
</body>
<footer>
<p class="copyright" id="copyright">
&copy; 2022
<span class="gradient-text">
DavidZ
</span>.
Powered by <a href="http://hexo.io/" title="Hexo" target="_blank" rel="noopener">Hexo</a>
Theme
<span class="gradient-text">
<a href="https://github.com/TriDiamond/hexo-theme-obsidian" title="Obsidian" target="_blank" rel="noopener">Obsidian</a>
</span>
<small><a href="https://github.com/TriDiamond/hexo-theme-obsidian/blob/master/CHANGELOG.md" title="v1.4.7" target="_blank" rel="noopener">v1.4.7</a></small>
</br>
鲁ICP备
<span class="gradient-text">
<a href="https://beian.miit.gov.cn/" title="19008089号-1" target="_blank" rel="noopener">19008089号-1</a>
</span>
</p>
</footer>
<script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.7/MathJax.js?config=TeX-AMS-MML_HTMLorMML">
</script>
<script>
MathJax.Hub.Config({
"HTML-CSS": {
preferredFont: "TeX",
availableFonts: ["STIX", "TeX"],
linebreaks: {
automatic: true
},
EqnChunk: (MathJax.Hub.Browser.isMobile ? 10 : 50)
},
tex2jax: {
inlineMath: [
["$", "$"],
["\\(", "\\)"]
],
processEscapes: true,
ignoreClass: "tex2jax_ignore|dno",
skipTags: ['script', 'noscript', 'style', 'textarea', 'pre', 'code']
},
TeX: {
noUndefined: {
attributes: {
mathcolor: "red",
mathbackground: "#FFEEEE",
mathsize: "90%"
}
},
Macros: {
href: "{}"
}
},
messageStyle: "none"
});
</script>
<script>
function initialMathJax() {
MathJax.Hub.Queue(function () {
var all = MathJax.Hub.getAllJax(),
i;
// console.log(all);
for (i = 0; i < all.length; i += 1) {
all[i].SourceElement().parentNode.className += ' has-jax';
}
});
}
function reprocessMathJax() {
if (typeof MathJax !== 'undefined') {
MathJax.Hub.Queue(["Typeset", MathJax.Hub]);
}
}
</script>
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/gitalk@1.6.2/dist/gitalk.min.css">
<script src="//cdn.jsdelivr.net/npm/gitalk@1.6.2/dist/gitalk.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/jquery@3.4.1/dist/jquery.min.js"></script>
<script src="/js/plugin.js"></script>
<script src="/js/obsidian.js"></script>
<script src="/js/jquery.truncate.js"></script>
<script src="/js/search.js"></script>
<script src="//cdn.jsdelivr.net/npm/typed.js@2.0.10/lib/typed.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/blueimp-md5@2.12.0/js/md5.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/social-share.js@1.0.16/dist/js/social-share.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/codemirror@5.45.0/lib/codemirror.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/javascript/javascript.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/css/css.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/xml/xml.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/htmlmixed/htmlmixed.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/clike/clike.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/php/php.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/shell/shell.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/python/python.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/cmake/cmake.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/powershell/powershell.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/codemirror@5.45.0/mode/yaml/yaml.min.js"></script>
<script src="/js/busuanzi.min.js"></script>
<script>
$(document).ready(function () {
if ($('span[id^="busuanzi_"]').length) {
initialBusuanzi();
}
});
</script>
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/photoswipe@4.1.3/dist/photoswipe.min.css">
<link rel="stylesheet" href="//cdn.jsdelivr.net/npm/photoswipe@4.1.3/dist/default-skin/default-skin.min.css">
<script src="//cdn.jsdelivr.net/npm/photoswipe@4.1.3/dist/photoswipe.min.js"></script>
<script src="//cdn.jsdelivr.net/npm/photoswipe@4.1.3/dist/photoswipe-ui-default.min.js"></script>
<!-- Root element of PhotoSwipe. Must have class pswp. -->
<div class="pswp" tabindex="-1" role="dialog" aria-hidden="true">
<!-- Background of PhotoSwipe.
It's a separate element as animating opacity is faster than rgba(). -->
<div class="pswp__bg"></div>
<!-- Slides wrapper with overflow:hidden. -->
<div class="pswp__scroll-wrap">
<!-- Container that holds slides.
PhotoSwipe keeps only 3 of them in the DOM to save memory.
Don't modify these 3 pswp__item elements, data is added later on. -->
<div class="pswp__container">
<div class="pswp__item"></div>
<div class="pswp__item"></div>
<div class="pswp__item"></div>
</div>
<!-- Default (PhotoSwipeUI_Default) interface on top of sliding area. Can be changed. -->
<div class="pswp__ui pswp__ui--hidden">
<div class="pswp__top-bar">
<!-- Controls are self-explanatory. Order can be changed. -->
<div class="pswp__counter"></div>
<button class="pswp__button pswp__button--close" title="Close (Esc)"></button>
<button class="pswp__button pswp__button--share" title="Share"></button>
<button class="pswp__button pswp__button--fs" title="Toggle fullscreen"></button>
<button class="pswp__button pswp__button--zoom" title="Zoom in/out"></button>
<!-- Preloader demo http://codepen.io/dimsemenov/pen/yyBWoR -->
<!-- element will get class pswp__preloader--active when preloader is running -->
<div class="pswp__preloader">
<div class="pswp__preloader__icn">
<div class="pswp__preloader__cut">
<div class="pswp__preloader__donut"></div>
</div>
</div>
</div>
</div>
<div class="pswp__share-modal pswp__share-modal--hidden pswp__single-tap">
<div class="pswp__share-tooltip"></div>
</div>
<button class="pswp__button pswp__button--arrow--left" title="Previous (arrow left)">
</button>
<button class="pswp__button pswp__button--arrow--right" title="Next (arrow right)">
</button>
<div class="pswp__caption">
<div class="pswp__caption__center"></div>
</div>
</div>
</div>
</div>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="//www.googletagmanager.com/gtag/js?id=UA-157733505-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'UA-157733505-1');
</script>
<script>
function initialTyped() {
var typedTextEl = $('.typed-text');
if (typedTextEl && typedTextEl.length > 0) {
var typed = new Typed('.typed-text', {
strings: ['凡事都要留几分', '颜值是第一生产力'],
typeSpeed: 90,
loop: true,
loopCount: Infinity,
backSpeed: 20,
});
}
}
if ($('.article-header') && $('.article-header').length) {
$(document).ready(function () {
initialTyped();
});
}
</script>
<!-- 例:百度统计 --> <script> var _hmt = _hmt || []; (function() { var hm = document.createElement("script"); hm.src = "https://hm.baidu.com/hm.js?your_code"; var s = document.getElementsByTagName("script")[0]; s.parentNode.insertBefore(hm, s); })(); </script>
</html>