A brief demonstration that uses a convolutional neural network to recognize Magic the Gathering cards.
I'm embarking on a project to bring a cheap & effective trading card sorter to market. This is the second repository that I've shared tackling the card recognition software. See links below for earlier work and a demonstration video that accompanies this repository. This demo shows just how effective machine learning can be at solving a complex task like card recognition. There are many hurdles still to overcome before a similar implementation can be used at scale, yet I hope to show how promising these early results are for the project.
If you'd like to stay informed about this project, consider joining the email list here.
See the demo video going through this code here.
See prior demo using optical character recognition & image hashing here.