-
Notifications
You must be signed in to change notification settings - Fork 25
Open
Description
Thanks for your work on this algorithm and for making it available.
Have you considered modifying the implementation to use robust geometric predicates?
The hope would be that it would allow you to avoid the random perturbation of points by CH_NOISE_VAL.
The comments indicate that the perturbation is needed to avoid problems with coincident points. If I guarantee that I do not pass any coincident points to convhull, should it be guaranteed to work? (I.e. is it sensitive to other kinds of degeneracies -- co-linear points, or co-planar points?)
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels