poirsonSpatioChromatic

 Pattern-color separable spatial filters from Poirson & Wandell (1993)

 [lum, rg, by, positions] = poirsonSpatioChromatic([sampPerDeg =241],[dimension = 2])

  These filters are fitted spatial response, formed by taking the weighted
  sum of two or three Gaussians. 

 Parameters
   sampPerDeg: filter sampling resolution (default=241 samps/deg)
   dimension:  filter for 1D or 2D format (default = 2) 

 See the Spatial-CIELAB implementation for other related ideas.

 Examples:
  [lum, rg, by, x]  = poirsonSpatioChromatic([],1);
  plot(x,lum,'k-',x,rg,'r--',x,by,'b:'); xlabel('Position (deg)')
  grid on

  [lum, rg, by, x]  = poirsonSpatioChromatic(120,2);
  mesh(x,x,lum); xlabel('Position (deg)'); ylabel('Position (deg)')
  mesh(x,x,by);

  To compute the spatial MTF of these filters compute
  clf
  [lum, rg, by, x]  = poirsonSpatioChromatic(241,1);
  lumMTF = abs(fft(lum)); plot(lumMTF)
  minFreq = 1/(max(x) - min(x));
  freq = [1:length(x)]*minFreq;
  l = (freq < 60);
  plot(freq(l),lumMTF(l)); grid on

  For 2D plots some fftshifting is required
  [lum, rg, by, x]  = poirsonSpatioChromatic(241,2);
  rgMTF = ifftshift(abs(fft2(fftshift(rg)))); mesh(rgMTF)
  minFreq = 1/(max(x) - min(x));
  freq = [1:length(x)]*minFreq; freq = freq - (max(freq)/2);
  [X,Y] = meshgrid(freq,freq);
  clf; mesh(X,Y,rgMTF);
  set(gca,'xlim',[-40 40],'ylim',[-40 40]); 

 Copyright ImagEval Consultants, LLC, 2005.