Does different image parameters affect image registration?
    4 visualizaciones (últimos 30 días)
  
       Mostrar comentarios más antiguos
    
    Colleen
 el 3 de Oct. de 2013
  
    
    
    
    
    Comentada: Alex Taylor
    
 el 8 de Oct. de 2013
            I am currently trying to register 2, 3D CT images of a human tibia. I have implemented imregister to complete this task. When I register the right to left leg of the same individual from the same day, the registration is excellent. However, when I register the left leg to the left leg of the same individual 3 years later the registration is very poor. The CT collected a slice every 0.4mm but I noticed that the image spacing was different in the other 2 planes. One CT scan had image spacing of 0.156 x 0.156 x 0.4 while the other was 0.195 x 0.195 x 0.4. Could this explain the poor registration? Is there a way to specific the image spacing to improve the registration? Please let me know if my question is unclear or if you need more information/code/images. Thank you.
0 comentarios
Respuesta aceptada
  Alex Taylor
    
 el 4 de Oct. de 2013
        
      Editada: Alex Taylor
    
 el 4 de Oct. de 2013
  
      Hi Colleen, I would recommend you take a look at the following example that ships with the Image Processing Toolbox:
It shows how spatial referencing objects can be used to specify the resolution of your images in each dimension and improve 3-D registration results with imregister.
In your case, the first CT image you described would have a spatial referencing object defined as:
 Rct = imref3d(size(yourImage), 0.156,  0.156, 0.4);
Where yourImage is the variable that contains the first CT image.
Passing spatial referencing information can be very useful in getting imregister/imregtform to converge.
2 comentarios
  Alex Taylor
    
 el 8 de Oct. de 2013
				Collen,
I'm happy to hear that! Thanks for letting me know it worked out.
Más respuestas (1)
  Ashish Uthama
    
 el 3 de Oct. de 2013
        Difference in spatial resolution differece might explain the poor registration.
0 comentarios
Ver también
Categorías
				Más información sobre Geometric Transformation and Image Registration en Help Center y File Exchange.
			
	Productos
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!


