Hey AR Community,
we’re looking into using credit cards as markers, so we can judge the physical size of objects in AR.
(They are uniformly sized and readily available, compared to printable 3D markers)
On iOS we can use the Vision-Framework, which has VNDetectRectanglesRequest to detect rectangles, which works very fast.
The closest thing in ARFoundation/ARCore I have found is Image Tracking , which also has a demo in the ARFoundation samples . I feel like image tracking is more specific, I.e. even if we make a few photos of different credit cards with different colors and sizes, thats not the same as writing an ML model for a rectangle of a certain aspect ratio?