<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Posts on Barkey Wolf Consulting</title>
    <link>https://barkeywolf.consulting/posts/</link>
    <description>Recent content in Posts on Barkey Wolf Consulting</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Sun, 22 Jun 2025 00:00:00 +0000</lastBuildDate><atom:link href="https://barkeywolf.consulting/posts/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Hospital Shift Scheduling with OR-Tools</title>
      <link>https://barkeywolf.consulting/posts/hospital-scheduling/</link>
      <pubDate>Sun, 22 Jun 2025 00:00:00 +0000</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/hospital-scheduling/</guid>
      <description>Find the code for this post at https://github.com/jjhbw/hospital-schedule
Physician Scheduling Shift scheduling of doctors in a hospital is a notoriously complex puzzle. This process is often a manual, time-consuming task that can take days of a planner&amp;rsquo;s time. Some departments even let the doctors do it themselves, and time spent creating shift schedules is time not spent on patient care.
As a typical software nerd, I couldn&amp;rsquo;t resist grossly underestimating this problem.</description>
    </item>
    
    <item>
      <title>Object tracking in video using a MOSSE tracker implemented in Rust</title>
      <link>https://barkeywolf.consulting/posts/mosse-tracker/</link>
      <pubDate>Sun, 01 Dec 2019 15:30:00 +0200</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/mosse-tracker/</guid>
      <description>Thanks to @chriamue for building this browser-based demo using wasm-pack!
Tracking objects in videos is a typical problem in computer vision. Note that tracking is distinct from detection: we&amp;rsquo;re assuming we&amp;rsquo;ve already detected the object we want to track (by finding its centroid or bounding box). When working on tracking, we want to find the object of interest again in subsequent video frames.
Numerous tracking algorithms exist. For a nice overview, check out the 2010 paper &amp;lsquo;Visual Object Tracking using Adaptive Correlation Filters by David S.</description>
    </item>
    
    <item>
      <title>Ahead-of-time compilation of a Tensorflow model for lightweight inclusion in a Rust program</title>
      <link>https://barkeywolf.consulting/posts/tf-aot-rust/</link>
      <pubDate>Sat, 09 Mar 2019 13:30:00 +0000</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/tf-aot-rust/</guid>
      <description>Why? Tensorflow&amp;rsquo;s new Accelerated Linear Algebra (XLA) framework comes with a lot of advantages, like Just In Time (JIT) compilation of computation graphs leading to speedups during training and inference on both GPU and CPU. Moreover, it supports Ahead-of-time (AOT) compilation as well through the tfcompile tool.
For me, AOT compilation is where it gets interesting. I&amp;rsquo;ve struggled to find a decent, lightweight way to put Tensorflow models in production on CPU-only systems.</description>
    </item>
    
    <item>
      <title>Building a minimalistic time-sorted Event Store using a key-value database in Go</title>
      <link>https://barkeywolf.consulting/posts/badger-event-store/</link>
      <pubDate>Sat, 22 Sep 2018 13:31:00 +0200</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/badger-event-store/</guid>
      <description>Event sourcing Lately, I&amp;rsquo;ve been looking into the Event Sourcing pattern as a way to make a simple auditable data storage system.
In a traditional database system, state is often stored as a collection of related tables or as documents. You can query this state and mutate it at will, but problems arise when you want to find out something about previous states of your database. This requires keeping track of changes to tables or documents which often inspires hacky solutions.</description>
    </item>
    
    <item>
      <title>Computer vision: video background subtraction using a Gaussian Mixture Model in Rust</title>
      <link>https://barkeywolf.consulting/posts/background-subtraction/</link>
      <pubDate>Tue, 17 Jul 2018 11:10:59 +0200</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/background-subtraction/</guid>
      <description>To get more of a feel for computer vision techniques and their applications, I decided to dive into one of computer vision&amp;rsquo;s most described applications: traffic monitoring. I wanted to not use OpenCV for once, and decided this was a great excuse to explore Rust, a modern systems language.
The full implementation discussed in this post can be found on Github.
Goal Count traffic in a fixed-view, color surveillance camera feed.</description>
    </item>
    
    <item>
      <title>Using the ZBar barcode scanning suite in the browser with WebAssembly</title>
      <link>https://barkeywolf.consulting/posts/barcode-scanner-webassembly/</link>
      <pubDate>Fri, 25 May 2018 00:00:00 +0000</pubDate>
      
      <guid>https://barkeywolf.consulting/posts/barcode-scanner-webassembly/</guid>
      <description>I don&amp;rsquo;t have time for your dry prose, show me the code!
 GitHub repo with the code in-browser demo  If you find any mistakes in the code or in this post, be sure to let me know! You can create an issue in the github repo or send me an email.
Goal So I wanted to make a cross-platform barcode scanner for a research project involving inventory management. Ideally, it would end up being a simple library that could be used in the browser of a mid-tier smartphone.</description>
    </item>
    
  </channel>
</rss>
