<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Egocentric on Infinite Script</title><link>https://www.infinitescript.com/tags/egocentric/</link><description>Recent content in Egocentric on Infinite Script</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Thu, 02 Jul 2026 23:20:00 +0900</lastBuildDate><atom:link href="https://www.infinitescript.com/tags/egocentric/index.xml" rel="self" type="application/rss+xml"/><item><title>Always Moving: 6D Poses for a Wearable Fisheye Rig</title><link>https://www.infinitescript.com/2026/07/always-moving-6d-poses-for-a-wearable-fisheye-rig/</link><pubDate>Thu, 02 Jul 2026 23:20:00 +0900</pubDate><guid>https://www.infinitescript.com/2026/07/always-moving-6d-poses-for-a-wearable-fisheye-rig/</guid><description>&lt;p&gt;In the &lt;a href="https://www.infinitescript.com/2026/06/the-missing-link-calibrating-a-sparse-rgb-rig-with-optitrack/"&gt;last post&lt;/a&gt; I calibrated eight ZED cameras bolted to a room&amp;rsquo;s walls so OptiTrack&amp;rsquo;s 3D positions could be projected onto each image. Those cameras never moved: one 6-DOF pose per camera is the whole answer.&lt;/p&gt;&#10;&lt;p&gt;This post is the opposite case: a rig &lt;strong&gt;worn on a person, moving every frame&lt;/strong&gt;. The goal is the same, but &amp;ldquo;the pose&amp;rdquo; is now a trajectory recovered at 20 Hz for the entire take, which changes which tools work. The rig is &lt;strong&gt;&lt;a href="https://github.com/Ropedia/HOMIE-toolkit"&gt;Homie&lt;/a&gt;&lt;/strong&gt;, &lt;a href="https://ropedia.com/blog/20251216_introducing_ropedia"&gt;Ropedia&lt;/a&gt;&amp;rsquo;s four-camera egocentric setup for recording manipulation data: a person wears it, does a task with their hands, and we want every tracked object and finger marker to land on the right pixel.&lt;/p&gt;</description></item></channel></rss>