frameDist {EBImage}R Documentation

Calculate pairwise distances between image frames or frames of two images

Description

Given two images with multiple frames, frameDist calculates a matrix of distances between frames (the more similar are the images, the smaller are the distances). Prerequisite: images should be centered and rotationally and scale aligned. The function is recommended with stackObjects with arguments rotate=combine=TRUE.

Usage

  ## S4 method for signature 'Image, Image':
  frameDist(x, y, r, g, b, blur=TRUE, 
      method="dist", verbose, ...)
  ## S4 method for signature 'Image, missing':
  frameDist(x, y, r, g, b, blur=TRUE, 
      method="dist", verbose, ...)

Arguments

x,y A image stacks.
r,b,g If x and y are in TrueColor mode, these values specify weights of red, green and blue channels in the resulting distance.
blur A logical indicating whether frames ought to be blurred before comparison (TRUE recommended).
method Method to use: dist - Euclidian distance, dot - dot product.
verbose Provides additional output as the function can be lengthy.
... Reserved.

Details

For Grayscale images, the distance of each pair of frames, xf and yf, is calculated as mean(abs(xf-yf)), where mean is taken over the pixels with at least xf or yf being non-zero. For TrueColor images, the distance is calculated as mean(sqrt(r*rc^2+g*gc^2+b*bc^2)), where rc,gc,bc are [0,1] ranged values of red, green and blue channels and r,g,b are the weights. Again, the mean is taken only over non-zero pixels.

Value

A matrix of distances. If y is missing - a square symmetric matrix for the distances between frames in the same image with diagonal elements set at Inf.

Author(s)

Oleg Sklyar, osklyar@ebi.ac.uk, 2007

See Also

stackObjects


[Package EBImage version 2.6.0 Index]