Our algorithm essentially does:
for each step
for each object of the scene
d = min(d, distance(object))
A possible optimisation is to instead swap the loops, allowing to evaluate first the objects more likely to be close, and stop evaluating expensive background objects as soon as possible.
Our algorithm essentially does:
A possible optimisation is to instead swap the loops, allowing to evaluate first the objects more likely to be close, and stop evaluating expensive background objects as soon as possible.