smooth.2dbasis: Construct a 2d-functional data object by smoothing data.

In haghbinh/fda2d: 2D Functional Data Analysis

Description

Discrete observations on one image are fit with a set of smooth curves, each defined by an expansion in terms of user-selected basis functions. The fitting criterion is least squares.

Usage

smooth.2dbasis(s, u, y, e, f)

Arguments

s	a set of argument values corresponding to the observations in the rows of the matrix y.
u	a set of argument values corresponding to the observations in the columns of the matrix y.
y	the matrix of observed values corresponding s nad u.
e	a functional data basis object for the first argument s of the bivariate function.
f	a functional data basis object for the second argument t of the bivariate function.

Value

A 2d-functional object of class 'bifd'.

Examples

require(imager)
im <- as.cimg(function(x,y)
sin(x/16)+cos(y/16),128,128)
plot(im)
y <- im %>% as.matrix
m <- nrow(y)
n <- ncol(y)
p<-9
q<-9
e <- create.bspline.basis(c(0,1),p)
f <- create.bspline.basis(c(0,1),q)
u <- seq(0,1,length.out= n)
s <- seq(0,1,length.out= m)
x <- smooth.2dbasis(s,u,y,e,f)
y_hat <- eval.bifd(s,u,x)
sqrt(sum((y-y_hat)^2))
im_hat <- as.cimg(y_hat)
par(mfrow=c(1,2))
plot(im,main="True image")
plot(im_hat,main="Smoothed image")

haghbinh/fda2d documentation built on May 7, 2019, 6:58 p.m.