<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: markusMB</title><link>https://news.ycombinator.com/user?id=markusMB</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 17 Jun 2026 08:16:23 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=markusMB" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by markusMB in "OpenCV 5 Is Here: The Biggest Leap in Years for Computer Vision"]]></title><description><![CDATA[
<p>Done a few projects with OpenCV myself, and your list of issues reads as if you throw OpenCV and opencv_contrib into the same bucket. Which you shouldnt. And maybe your assessment is outdated here and there and it is time to look again.<p>- OpenCV is Apache license. Yes, it used to be more complicated.<p>- The only patented algorithm I am aware of, SIFT, used to be part of opencv_contrib. And the README in opencv_contrib would greet you with a warning, that the code may not be fit commercial use for various reasons. Only when the patent expired, it was moved into OpenCV core.<p>- Same observation for Aruco marker detection, which was in contrib for a long time because the options to choose from were either not-well-maintained or GPL-licensed code. It is now in core OpenCV (and Apache).<p>- Despite its age, I think that OpenCV is still more than relevant today. And being part of modern languages like C++, Swig, Java and Python (and for years already) is part of that. Still I was surprised how long they maintained OpenCV 2 and 3.<p>- Over the past releases and few years, my impression was actually that core API was very much stable(izing). Cant say what happened in contrib – or what it feels like when you treat core and contribute as one and a feature progressed from contributing to core.<p>- I do agree, that I usually I would check that a MINOR releases wasnt actually a MAJOR release, breaking some API or behavior I was relying on. I am hoping that Version 5 is pulling the ambitions for making things differently away from Version 4. So v4 can be used stably ;-)</p>
]]></description><pubDate>Tue, 09 Jun 2026 16:44:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=48463541</link><dc:creator>markusMB</dc:creator><comments>https://news.ycombinator.com/item?id=48463541</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48463541</guid></item><item><title><![CDATA[New comment by markusMB in "Playing with Vision Embeddings"]]></title><description><![CDATA[
<p>Beautiful illustrations  
I find, 'Playing' is just the free and motivated version of 'exploration'.<p>One thought on your nicely illustrated "key observation [is] that neural networks tend to place features along directions": my guess is that the neural net was TOLD to behave that way by choosing e.g. Cosine Loss?</p>
]]></description><pubDate>Mon, 08 Jun 2026 09:06:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=48442941</link><dc:creator>markusMB</dc:creator><comments>https://news.ycombinator.com/item?id=48442941</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48442941</guid></item><item><title><![CDATA[Ask HN: Wishlist for an Immich-macOS-App?]]></title><description><![CDATA[
<p>Thanks to HN coverage here, I learned about Immich, got curious, tried it, became excited about it. The other day, I wanted to share my excitement with a non-tech relative – but didnt really dare to suggest how simple it is to get going. I am sure I would have lost him way before "Docker", "Linux", or "Hetzner".<p>So I thought I put together an app that 1. runs Immich using Apple's container VMs, 2. takes care of Immich's containers, 3. opens Immich in a local browser at the click of a button, 4. displays its current IP address in order to connect with your smartphone, 5. handles configs and updates (restarts containers as needed).<p>What features would you suggest or would be important for you? My personal items on the list:<p>- single data directory containing image data as well as database backups<p>- data directory on an external drive: it may not be connected or be ejected while containers are running<p>- current IP address as a QR code<p>- ...?</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48311064">https://news.ycombinator.com/item?id=48311064</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Thu, 28 May 2026 16:15:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48311064</link><dc:creator>markusMB</dc:creator><comments>https://news.ycombinator.com/item?id=48311064</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48311064</guid></item><item><title><![CDATA[New comment by markusMB in "5x5 Pixel font for tiny screens"]]></title><description><![CDATA[
<p>Turns out, just by adding a 1 pixel border, you can (visually) turn the characters into a Aruco AR code dictionary :-) The dictionary created from the whole font does have uniqueness collisions – it depends on your use case if that is problematic.
But you can create a font of AR codes that can be read by machine as well as human:
<a href="https://codeberg.org/markusBM/ARuChars#readme" rel="nofollow">https://codeberg.org/markusBM/ARuChars#readme</a></p>
]]></description><pubDate>Sun, 03 May 2026 19:18:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=48000359</link><dc:creator>markusMB</dc:creator><comments>https://news.ycombinator.com/item?id=48000359</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48000359</guid></item></channel></rss>