TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Graph2Plan: Learning Floorplan Generation from Layout Graphs

55 点作者 nonoesp大约 5 年前

5 条评论

carapace大约 5 年前
Good, now teach it Pattern Language.<p>Christopher Alexander FTW!<p><a href="https:&#x2F;&#x2F;www.livingneighborhoods.org&#x2F;ht-0&#x2F;bln-exp.htm" rel="nofollow">https:&#x2F;&#x2F;www.livingneighborhoods.org&#x2F;ht-0&#x2F;bln-exp.htm</a>
评论 #23169934 未加载
charleskinbote大约 5 年前
Reminds me of similar work done with genetic algorithms: <a href="https:&#x2F;&#x2F;www.joelsimon.net&#x2F;evo_floorplans.html" rel="nofollow">https:&#x2F;&#x2F;www.joelsimon.net&#x2F;evo_floorplans.html</a>
nrjames大约 5 年前
Also reminds me of Joris Dorman&#x27;s cyclic dungeon work for the game Unexplored: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=_wvkTT-6P3Q" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=_wvkTT-6P3Q</a>
nonoesp大约 5 年前
A similar model is #HouseGAN done by researchers Nelson Nauata &amp; Chin-Yi Cheng at Autodesk Research in 2019.<p>Post → <a href="https:&#x2F;&#x2F;nono.ma&#x2F;housegan" rel="nofollow">https:&#x2F;&#x2F;nono.ma&#x2F;housegan</a> Paper → <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2003.06988" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2003.06988</a> GIF → <a href="https:&#x2F;&#x2F;nono.imgix.net&#x2F;img&#x2F;u&#x2F;post-housegan-200401.gif?ixlib=php-3.1.0" rel="nofollow">https:&#x2F;&#x2F;nono.imgix.net&#x2F;img&#x2F;u&#x2F;post-housegan-200401.gif?ixlib=...</a>
vlovich123大约 5 年前
Could be cool to expand this to electoral map building. Then you don&#x27;t even need to manually draw the lines. Publish the loose constraints &amp; the source data &amp; everyone can regenerate their own electoral map to confirm no funny business.