I am currently working on an interest region extraction algorithm based on the work of Ranjith Unnikrishnan (you can see the current state in features/statistical_multiscale_interest_region_extraction.h. For this, I need geodesic distances between any point pair in a cloud. Furthermore, I also need range searches using the geodesic distance metric.
For this, I just used Boost Graph Library to create a K-nearest neighbor graph (using FLANN) and then applying the Johnson algorithm to get the shortest path between every pair (recommended for sparse graphs). The problem I have now is in the range searches - I see no simple/direct solution except doing some more expensive graph searches.
Did anyone work on similar things and could share a few hints?
This post has NOT been accepted by the mailing list yet.
I am working on geodesic distnace computation between pair of points on PCD. If you have found solution for your work, can you share the code or clear steps for implementation using Johnson algorithm.
Also whats the solution for dense graphs?