GMM_Wd: GMM_Wd

View source: R/registration.R

GMM_WdR Documentation

GMM_Wd

Description

Compute 2-Wasserstein distance between two Gaussian mixture models See: Delon J, Desolneux A. (2019) A Wasserstein-type distance in the space of Gaussian Mixture Models. hal-02178204v2

Usage

GMM_Wd(m1, m2, S1, S2, w1 = NULL, w2 = NULL, S = NULL)

Arguments

m1

matrix of means of first GMM

m2

matrix of means of second GMM

S1

array of covariance matrices of first GMM such that m1[i,] has covariance matrix S1[,,i]

S2

array of covariance matrices of second GMM such that m2[i,] has covariance matrix S2[,,i]

w1

(optional) vector of mixture weights of first GMM.

w2

(optional) vector of mixture weights of second GMM.

S

(optional) array of pre-computed sqrtm(sqrtm(S1[,,i]) %*% S2[,,j] %*% sqrtm(S1[,,i]))

Value

list of distance value d and optimal transport matrix ot


LOMAR documentation built on March 18, 2022, 6:05 p.m.