How is marker recognition done in ar.js? How can we improve recognition efficiency?

Question

I want to know how exactly the marker is generated/recognized so that I can make it more efficient in my code. Is there any way of making something a marker other than a patt file?

I was trying to use alphabets as markers and when displaying many letter-markers in one frame, it fails to detect the right one. So I wanted to make any random, clear image with the black border to be recognized.

Answer 1

You can use Jerome's maker generator to produce a marker of any image, but the image should be rotationally assymetric as per the ARToolKit docs