Hands-on Tutorials

Using one-shot learning to identify thousands of classes with very few images per class.

There is a childhood card game that I have liked when I was younger called Yugioh. It involves summoning different monsters as well as using different spells and traps with the goal of destroying your opponent’s monsters and reducing their life points to 0.

In this game as of right now, there are over 10,000 cards in the game (10,856 to be exact). And I decided to take the ambitious task of making an AI to recognize all of them.

Results on Yugioh cards in the wild (image by author)

Got inspiration to try this out after seeing when Konami made:

https://ygorganization.com/konami-staff-develop-powerful-image-recognition-technology/

I wanted to try to replicate as close…

Anthony Lowhur

Full-time CV and ML engineer | AI and Deep Learning enthusiast https://www.linkedin.com/in/anthony-lowhur-886264b1 | http://vanstorm9.github.io/

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store