I am reading the following paper: http://research.micr...perfecthash.pdf
I think the idea it presents is summarised by the figures at the top of the paper. I think the idea is a way of mapping non uniformly spatially distributed data to a uniformly distributed minimal hash table.
I think it does this by using a hash function of the form: h(p) = h0(p) + O[h1(p)] , where h is an index in the hash table, p(x,y,z) are the spatial coordinates of the data point we wish to hash, and O is an offset table smaller than the hash table itself.
But that's as far as my "understanding" of the paper goes.
They keep saying how simple their method is but I don't find their explanation of how to actually implement this simple at all. What are the functions for h0 and h1? How do we calculate the size of h and O and the offset values for O?
Perfect and Minimal Spatial Hashing
1 reply to this topic
1 user(s) are reading this topic
0 members, 1 guests, 0 anonymous users