[Insight-users] registration for tiled volume

[Luis Ibanez]

Hi Rocky,

The procedure that you posted describes pretty well what
you should do for merging the two images in a single one.

In your case, you will have to use iterators explicitly
since you have to do the blending as you copy the pixel
values into the composite image.

There are other filters that could do blending, but not of
the type that you need. Stitching is very specific in the
sense that the blending coefficients vary over the image.

Again, the solution for the long term involves to implement
classes specifically designed for supporting stitching.
That should include a class for performing resampling and
probably incorporating blending on the overlaping regions.

The iterator loops that you are going to write will end up
being the content of one of those new classes.

Please let us know if you have further questions,


     Thanks


        Luis


---------------------
Rocky Rhodes wrote:
>>One way to actually stitch the images together is to create an image
>>large enough for containing the group of images that you want see
>>together in a single file, and then use the itkResampleImageFilter,
>>in order to actually move the intensities to that destination image.
>>
> 
> 
> To be clear, is this something like what I need to do:
> 
> 	1. Have readers for both images
> 	2. Use itkResampleImageFilter to transform moving image
> appropriately.
> 	3. Create a new image and allocate with size appropriate for the
> combined image.
> 	4. Iterate over pixels in first image and combined image, Set'ing
> pixels in combined image to values returned by Get for first image.  Use
> region from first image for these iterators.
> 	5. Repeat for second image, now using different region reflecting
> the footprint of the second image.
> 	6. Write out combined image.
> 
> Or is there a way to copy a region of pixels from one image to another?
> 
> 
>>The blending policy is arguable. It is common to do it as a linear
>>blending where the weights are related to how close the pixel is
>>from the non-overlapping border of each image. You may have to
>>experiment a bit on this front, since it depends a lot on how similar
>>the two overlapping image are (e.g. hopping that there are no
>>optical distortions or illumination artifacts. Modern microscopes are
>>quite good regarding aberration elimination, so you probably will not
>>have any issues of that sort.
>>
> 
> 
> 
> To do blending, do I need to explicitly perform the linear blending in the
> loops above?  Again, is there a way to blend two images or regions from two
> images together without the iterator loops?
> 
> Thanks for the help.
> 
> 	Rocky
> 
>