Personally , I find the problem has more to do with the way you shuffle your cards as it makes it a bit more difficult to identify what card is what.
Let's say for instance that instead of shuffling your embed assets , you were first creating your cards , naming them after the image they add , then creating an Array ( or Vector ) of cards and shuffle that Array ( or Vector ), identifying the cards would be fairly easy since you'd only have to compare their name property.