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。

Direct Voxel Grid Optimization

78 点作者 lnyan将近 3 年前

4 条评论

momojo将近 3 年前
I realize the Plenoxels paper is already mentioned by the author, but 2 Minute Papers does a great overview of the concepts (for those wondering what makes this different from other techniques)<p>Here&#x27;s the NeRF paper (2020) <a href="https:&#x2F;&#x2F;www.matthewtancik.com&#x2F;nerf" rel="nofollow">https:&#x2F;&#x2F;www.matthewtancik.com&#x2F;nerf</a><p>And the more recent Plenoxels explainer (2022) <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=yptwRRpPEBM" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=yptwRRpPEBM</a><p>The original NeRF&#x27;s novelty was the neural-net used to extrapolate the final images, but the newer paper shows that that&#x27;s not really necessary. My understanding is that the secret-sauce is the 5D &#x27;plenoxel&#x27;.
chroem-将近 3 年前
What&#x27;s the difference from regular photogrammetry?
评论 #31745773 未加载
评论 #31747563 未加载
tehsauce将近 3 年前
On first glance it appears older methods like TensoRF and nvidias instant-nerf are far superior to this.
评论 #31748463 未加载
mysterEFrank将近 3 年前
I don&#x27;t get it, the point of NeRF is to replace a voxel grid with an MLP
评论 #31748355 未加载