Michael Gold

One of the most frequent complaints I hear about developing for the Quest is: I have to confirm the guardian every time I put on my headset – and this can be many times during a coding session. If you’ve always wanted to disable these notifications, you’re in luck. With the latest firmware installed on […]

This past May, Google announced an open source computer vision and deep learning project called PoseNet for TensorFlow.js that is able to estimate human poses using only Javascript. I decided to check out the library and see what type of results I could get using the webcam on my laptop. Open Pose, the previous real-time

This is officially our 200th daily post, and a good time for an announcement about what’s to come on our website. We’ve upgraded our server to a newer version of PHP.  This means we’ll be able we were able to configure a modern WordPress theme to organize all of the content on the site. So

Over the past few years, I have put together a software stack for storing, organizing, and annotating the PDF documents – mainly slides and research papers. I wanted a system that would let me take notes on my documents from any of my devices and would give me the ability to search through all my

One of the nicest things about using vim as a text editor is that the key bindings allow you to control the entire program with very few movements away from the home row of keys. The main exception to this rule is that in order to exit out of text insertion mode, you need to

In his 1889 essay “The Decay of Lying,” Oscar Wilde called for the return to “lying” — not as politicians lie by misrepresenting facts, but as artists lie through exaggeration. Bored by the realism art movement of his time, Wilde argued, exaggerations are what makes art aspirational, and, “Life imitates Art far more than Art

One of the things I like most about studying a new topic is learning about the tools used by practitioners in that field. Through my process of implementing computer vision / deep learning research papers, I became aware of the Python Pandas module that is widely used by data scientists. When used within a Jupyter

At our weekly lab meeting this past Friday, our research group conducted an exercise. We each spoke about examples from a possible not-to-distant future where screens had been replaced with less obtrusive technology. Our goal in this process was to call out the things that had changed. There was only one rule: Technology changed things

As I continue my research in deep learning and computer vision, my goal is to learn more about the limitations of what is currently one of the most popular methods for estimating human poses – CMU’s OpenPose system. As shown in the video above, this method can produce impressive results when the entire body of the subject

Today, I implemented the OpenPose library on two of my machines. OpenPose uses deep learning through convolutional networks to estimate human poses from single camera footage. My first attempt – implementing OpenPose on my laptop For this first attempt, I compiled OpenPose on my laptop which is running Ubuntu 18.04 and has an integrated Intel