tpsGetWarp

PURPOSE ^

Given two sets of corresponding points, calculates warp between them.

SYNOPSIS ^

function [warp,L,LnInv,bendE] = tpsGetWarp( lambda, xsS, ysS, xsD, ysD )

DESCRIPTION ^

 Given two sets of corresponding points, calculates warp between them.

 Uses booksteins PAMI89 method.  Can then apply warp to a new set of
 points (tpsInterpolate), or even an image (tpsInterpolateIm).
  "Principal Warps: Thin-Plate Splines and the Decomposition of
  Deformations".  Bookstein.  PAMI 1989.

 USAGE
  [warp,L,LnInv,bendE] = tpsGetWarp( lambda, xsS, ysS, xsD, ysD )

 INPUTS
  lambda      - rigidity of warp (inf means warp becomes affine)
  xsS, ysS    - [1xn] correspondence points from source image
  xsD, ysD    - [1xn] correspondence points from destination image

 OUTPUTS
  warp        - bookstein warping parameters
  L, LnInv    - see bookstein
  bendE       - bending energy

 EXAMPLE - 1
  xsS=[0 -1 0 1];  ysS=[1 0 -1 0];  xsD=xsS;  ysD=[3/4 1/4 -5/4 1/4];
  warp = tpsGetWarp( 0, xsS, ysS, xsD, ysD );
  [gxs, gys] = meshgrid(-1.25:.25:1.25,-1.25:.25:1.25);
  tpsInterpolate( warp, gxs, gys, 1 );

 EXAMPLE - 2
  xsS = [3.6929 6.5827 6.7756 4.8189 5.6969];
  ysS = [10.3819 8.8386 12.0866 11.2047 10.0748];
  xsD = [3.9724 6.6969 6.5394 5.4016 5.7756];
  ysD = [6.5354 4.1181 7.2362 6.4528 5.1142];
  warp = tpsGetWarp( 0, xsS, ysS, xsD, ysD );
  [gxs, gys] = meshgrid(3.5:.25:7, 8.5:.25: 12.5);
  tpsInterpolate( warp, gxs, gys, 1 );

 See also TPSINTERPOLATE, TPSINTERPOLATEIM, TPSRANDOM

 Piotr's Computer Vision Matlab Toolbox      Version 2.0
 Copyright 2014 Piotr Dollar.  [pdollar-at-gmail.com]
 Licensed under the Simplified BSD License [see external/bsd.txt]

Generated by m2html © 2003