minimum_contrast_images: Minimum contrast method estimate image correations

In kangjian2016/SpatCorrImage: Spatial Adaptive Varying Corraltion Analysis for for Multimodal Neuroimaging Data

Description

This function estimate the rho parameter in the exponetial squared covariance kernel exp(-rho*||x - x'||^2)

Usage

minimum_contrast_images(X, coords)

Arguments

X	a matrix format of multiple images, where rows are different images in vector format.
coords	voxel locations of images: each raw has the (x,y) coordinates for 2D images or (x, y, z) coordinates for 3D images.

Value

the rho value

Author(s)

Jian Kang <[e-mail]>

Examples

n = 10
coords = expand.grid(seq(-1,1,length=n),seq(-1,1,length=n))
rho0 = 0.5
dist_coords = dist(coords)
Sigma = exp(-rho0*as.matrix(dist_coords^1.99))
R <- chol(Sigma)
m = nrow(coords)
N = 20
X <- t(crossprod(R,matrix(rnorm(m*N),nrow=m,ncol=N)))
image(matrix(X[1,],n,n))
minimum_contrast_images(X,coords)

kangjian2016/SpatCorrImage documentation built on Oct. 22, 2021, 1:21 a.m.