Hi Steve,
Thanks so much for your quick response.
The data type of the realigned phase images is INT16. We've also tried
unwrapping the phase images before realignment, and in that case the
realigned unwrapped phase images are FLOAT32, although after realignment
the same images are blank, (showing a signal intensity of zero throughout the
image.)
Phase unwrapping with prelude works fine for the whole time series, and the
only preprocessing we've done is to convert the images from dicom to Analyze
with medcon. There doesn't appear to be anything unusual about the mat
files associated with the images which are blank after realignment. (We've
tried realigning those particular images with other mat files which were applied
successfully to other images in the same time series, but those images are still
blank after realignment so the problem appears to be with the images rather
than the transformations.)
On our scanner the phase images are saved as 16-bit integer dicom
files where the signal intensity is multiplied by a scaling factor to increase the
contrast for display purposes. I was wondering if the problem might be related
to the range of pixel values, but when we run prelude we typically divide by
the scale factor and save the images as floats beforehand, and since the
same images appear blank after realignment and phase unwrapping (with a
narrower range of pixel values, and 32 bit images) it doesn't look like the range
of pixel values is the problem.
Any ideas/suggestions would be very welcome!
Thanks very much,
Ruth
|