Algorithm 1 NFT-based pose estimation

input: s = scene     c = camera     r = renderer    imdes = descriptor/pattern 1: Ar = ARsession (s,c,r) 2:  for trackingbackend do 3:      onRender.push(Ar, imdes,) 4:        if imdes > tolerance 5:         keyframe = calculateKeyframe [R|t], 6:         anchor = createAnchor(x, y); 7:        else 8:         continue trackingbackend 9:   end for 10:    GLTransformation (x, y, keyframe)  output: normalized direction vector