opencv - R function Mclust slow -
i used mclust function in mclust package em-clustering of vector of 27,000 entries 2 clusters:
mclust(data_vector, g=2)
another software uses opencv em-clustering 3 times faster mclust (even if reduce maximum number of iterations in mclust e.g. 4). in mclust source looks function implemented in fortran.
how can it seems slower opencv implementation?
try running both exact same:
- initial conditions
- model (with/without covariance etc.)
i believe mclust
quite expensive initialization. if opencv starts random sample initialization, no wonder faster.
so starter, give both exact same vector start with.
Comments
Post a Comment