<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Rendering Natural Camera Bokeh Effect with Deep Learning · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2006.05698/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2006.05698/feed.xml" rel="self" type="application/rss+xml" /><item><title>Rendering Natural Camera Bokeh Effect with Deep Learning — mentioned by skylanh</title><link>https://news.ycombinator.com/item?id=25668300</link><guid isPermaLink="false">urn:hackerpapers:mention:2006.05698:25668300</guid><pubDate>Thu, 07 Jan 2021 07:37:33 GMT</pubDate><description>&lt;p&gt;Depth of field effects caused by long focal length lenses (preferred in portraiture for this reason), called &amp;quot;bokeh&amp;quot;. In this case the light is coming into front lens elements, and it doesn&amp;#x27;t appear to be corrected for distortion, and then not uniformly focused onto the surface of the camera&amp;#x27;s sensor. Potentially this was for a Imax dome. So, no, this doesn&amp;#x27;t have anything to do with smart phones. In ye&amp;#x27; olde times, induced background blur would have been a gaussian blur in software. I&amp;#x27;m not sure what special magic they do today. https://arxiv.org/abs/2006.05698 suggests things have moved on.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2006.05698/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CV</category></item></channel></rss>
