Help for package cmR

Type:

Package

Title:

Analysis of Cardiac Magnetic Resonance Images

Version:

1.1

Date:

2023-07-12

Author:

Volker Schmid [aut, cre]

Maintainer:

Volker Schmid <volker.schmid@lmu.de>

Depends:

R (≥ 3.5.0)

Imports:

Matrix, splines, fields, graphics, parallel, plotrix

Description:

Computes maximum response from Cardiac Magnetic Resonance Images using spatial and voxel wise spline based Bayesian model. This is an implementation of the methods described in Schmid (2011) <doi:10.1109/TMI.2011.2109733> "Voxel-Based Adaptive Spatio-Temporal Modelling of Perfusion Cardiovascular MRI". IEEE TMI 30(7) p. 1305 - 1313.

License:

GPL-3

RoxygenNote:

7.2.3

Encoding:

UTF-8

LazyData:

true

URL:

https://bioimaginggroup.github.io/cmr/

Suggests:

knitr, rmarkdown, codetools, testthat (≥ 3.0.0), R.rsp

VignetteBuilder:

knitr, R.rsp

BugReports:

https://github.com/bioimaginggroup/cmR/issues

NeedsCompilation:

Packaged:

2023-07-19 08:58:24 UTC; schmid

Repository:

CRAN

Date/Publication:

2023-07-19 10:30:02 UTC

Bullseye plot

Description

Bullseye plot

Usage

bullseye(x, lim = NULL, reverse = TRUE, legend = TRUE, text = TRUE, cex = 1)

Arguments

x

vector of length 16 or 17

lim

limits of x values

reverse

boolean, reverse colors?

legend

boolean, add legend?

text

boolean, should text legend be added?

cex

cex for text legend

Value

plot

Examples

  bullseye(1:16)

Bayesian analysis of cardiovascular magnetic resonance imaging

Description

Bayesian analysis of cardiovascular magnetic resonance imaging

Usage

cmr(
  data,
  input,
  mask = NULL,
  method = "spatial",
  quantiles = c(0.25, 0.75),
  cores = parallel::detectCores()
)

Arguments

data

3D or 4D array of CMR signal

input

input function

mask

2d array of mask. Voxel with 0 or FALSE will be omitted from analysis. Default NULL: use NA values in data as mask

method

"spatial" or "local"

quantiles

quantiles used for credible interval, default: c(0.25, 0.75)

cores

number of cores for parallel computation. Spatial model only computes slices parallel, local can be parallelized on voxel level

Value

list of mbf (point estimation) and ci (credible interval)

Spline analysis of cardiovascular magnetic resonance imaging

Description

Spline analysis of cardiovascular magnetic resonance imaging

Usage

cmr.local(data, mask, input, quantiles = c(0.25, 0.75), cores = 1)

Arguments

data

3d array of CMR signal

mask

2d array of mask. Voxel with 0 or FALSE will be omitted from analysis

input

input function

quantiles

quantiles used for credible interval, default: c(0.25, 0.75)

cores

number of cores to use in parallel computing

Value

list of mbf (point estimation) and ci (credible interval)

Examples


oldpar <- par(no.readonly = TRUE)
 library(cmR)
 data(cmrsim)
 local.mbf=local.ci=array(NA,c(30,30,3))
 for (i in 1:3){
  mask=array(NA,c(30,30))
  mask[cmrdata_sim[,,i,1]!=0]=1
  temp=cmr.local(cmrdata_sim[,,i,], mask, input_sim, cores=2)
  local.mbf[,,i]=t(as.matrix(temp$mbf))
  local.ci[,,i]=t(as.matrix(temp$ci))
 }
 par(mfrow=c(2,1))
 imageMBF(maxresp_sim, zlim=c(0,5))
 imageMBF(local.mbf, zlim=c(0,5))
 imageMBF(local.ci, zlim=c(0,0.8))
par(oldpar)

Spatial spline analysis of cardiovascular magnetic resonance imaging

Description

Spatial spline analysis of cardiovascular magnetic resonance imaging

Usage

cmr.space(data, mask, input, quantiles = c(0.25, 0.75))

Arguments

data

3d array of CMR signal

mask

2d array of mask. Voxel with 0 or FALSE will be omitted from analysis

input

input function

quantiles

quantiles used for credible interval, default: c(0.25, 0.75)

Value

list of mbf (point estimation) and ci (credible interval)

Examples


oldpar <- par(no.readonly = TRUE)
 library(cmR)
 data(cmrsim)
 mask=array(NA,c(30,30))
 space.mbf=space.ci=array(NA,c(30,30,3))
 for (i in 1:3){
  mask=array(NA,c(30,30))
  mask[cmrdata_sim[,,i,1]!=0]=1
  temp=cmr.space(cmrdata_sim[,,i,], mask, input_sim)
  space.mbf[,,i]=t(as.matrix(temp$mbf))
  space.ci[,,i]=t(as.matrix(temp$ci))
  }
 par(mfrow=c(2,1))
 imageMBF(maxresp_sim, zlim=c(0,5))
 imageMBF(space.mbf, zlim=c(0,5))
 imageMBF(space.ci, zlim=c(0,0.8))
 par(oldpar)

Simulated data for CMR package.

Description

This data set is provided as example for the usage of the cmR package. cmrdata_sim is a simulated CMR image.

Usage

cmrdata_sim

Format

A 4D array, 30x30 pixels for 3 slices at 30 time points.

Plotting of (voxelwise) cardiac MBF

Description

Plotting of (voxelwise) cardiac MBF

Usage

imageMBF(img, zlim = NULL, reverse = TRUE)

Arguments

img

3d array ob MBF values

zlim

limits of MBF, default: NULL means zlim=c(0,max(img,na.rm=TRUE))

reverse

reverse color scheme

Value

plots

Examples

data(cmrsim)
imageMBF(maxresp_sim)

Simulated data for CMR package.

Description

This data set is provided as example for the usage of the cmR package. input_sim is the simulated input function.

Usage

input_sim

Format

Vector for 30 time points.

Simulated data for CMR package.

Description

This data set is provided as example for the usage of the cmR package. maxresp_sim is the true maximum response used in the simulation.

Usage

maxresp_sim

Format

A 3D array, 30x30 pixels for 3 slices.

Pseudo bullseye plot

Description

Pseudo bullseye plot

Usage

pseudobullseye(
  x,
  lim = range(x, na.rm = TRUE),
  legend = FALSE,
  text = TRUE,
  reverse = FALSE,
  center = TRUE,
  cex = 1,
  legend.width = 1
)

Arguments

x

3D array

lim

limits of x values

legend

boolean, add legend?

text

boolean, should text legend be added?

reverse

boolean, reverse colors?

center

boolean, should input x be centered before plotting

cex

cex for text legend

legend.width

Width in characters of the legend strip.

Value

plots

Examples

data(cmrsim)
pseudobullseye(maxresp_sim)

Draw random vectors from multivariate Gaussian in canonical form

Description

Draw random vectors from multivariate Gaussian in canonical form

Usage

rmvnormcanon(n, b, P)

Arguments

n

Number of draws

b

b parameter

P

Precision matrix

Value

matrix with n columns, vector if n=1

Examples

P<-matrix(c(1,.5,.5,1),ncol=2)
b=c(2,0)
# expected value and covariance matrix
Sigma = solve(P)
mu = b%*%Sigma
# sample
x<-rmvnormcanon(1000,b,P)
mu.hat=apply(x,1,mean)
print(mu.hat-mu)
Sigma.hat=var(t(x))
print(Sigma.hat-Sigma)