Too Long; Didn't Read
This post tells the story of how I built an image classification system for <strong>Magic cards</strong> using <em>deep convolutional denoising autoencoders</em> trained in a <em>supervised</em> manner. There are <strong>tens of thousands different cards</strong>, many cards look <strong>almost identical</strong> and <strong>new cards are released several times a year</strong>. Combined, this makes for a <strong>highly interesting</strong> Machine <a href="https://hackernoon.com/tagged/learning" target="_blank">Learning</a> challenge.