imregister translation mode, suppress registration in second direction

Hello,

is possilble to suppress second shift direction in imregister function during 'tranlastion' mode? I have some images with moving feature but only in one direction, for example x direction, and need to do registration only in this direction. Usually it works good but sometimes there is also unwanted registration in y direction. Thanks.

Answer 1

[Matt J]

One hack that I can think of (and I emphasize that it is only a hack) is to deliberately corrupt the images with bright horizontal line patterns and apply imregtform. That way, a strong gradient against shifts in the y-direction will be introduced into the registration cost function. I illustrate this in the example below where the true translation is [10,0.6], but with this technique the registration is forced to reach a translation estimate in which the y-component is greatly suppressed. It should work even better, however, in your case where the true translation really does have y=0.

%% Simulation
moving = dicomread('knee1.dcm');
fixed=imtranslate(moving,[10,0.6]);
[optimizer, metric] = imregconfig('monomodal');



%% Corrupt the images with bright horizontal lines
Mcorr=moving; Mcorr(1:2:end,:)=3000; 
Fcorr=fixed; Fcorr(1:2:end,:)=3000;


tform0 = imregtform(moving, fixed,'translation',optimizer,metric);
tform = imregtform(Mcorr,Fcorr,'translation',optimizer,metric);

base_tform = tform0.T, 

constrained_tform = tform.T

The results are

base_tform =

    1.0000         0         0
         0    1.0000         0
   10.0000    0.6000    1.0000


constrained_tform =

    1.0000         0         0
         0    1.0000         0
    9.9149    0.0033    1.0000

[Matt J]

Here's another variation that uses the complete input images

%%corrupt the images
[mm,nn]=size(moving);Z=zeros(2*mm,nn)+4000;

Mcorr=Z; 
  Mcorr(1:2:end,:)=moving; 
Fcorr=Z; 
  Fcorr(1:2:end,:)=fixed; 

[Rene Riha]

Yeah, this is really smart hack, thanks! :D

[Rene Riha]

What I can do is also just simply put tform.T(3,2)=0, use the imwarp_same function and y shift disappears completely.

[Matt J]

You can do that and it might be adequate for you, but my suggested approach is much closer to actually limiting the tform model to a 1D translation.