Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote:
Hello,
We have been using mri_convert to convert our dicoms into .nii niftis. There is an entry in the .nii header for the slope scale, but mri_convert appears to discard this value instead of copying it into the header (whereas mricron retains this value) We prefer to use mri_convert.
Is there a way to have mri_convert carry this value over? Why is it being discarded?
Thanks, Eli
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote:
Hello,
We have been using mri_convert to convert our dicoms into .nii niftis. There is an entry in the .nii header for the slope scale, but mri_convert appears to discard this value instead of copying it into the header (whereas mricron retains this value) We prefer to use mri_convert.
Is there a way to have mri_convert carry this value over? Why is it being discarded?
Thanks, Eli
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote:
Hello,
We have been using mri_convert to convert our dicoms into .nii niftis. There is an entry in the .nii header for the slope scale, but mri_convert appears to discard this value instead of copying it into the header (whereas mricron retains this value) We prefer to use mri_convert.
Is there a way to have mri_convert carry this value over? Why is it being discarded?
Thanks, Eli
What is the dicom tag for the scale?
On 11/10/2016 01:29 PM, Rockers, Elijah D. wrote:
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote:
Hello,
We have been using mri_convert to convert our dicoms into .nii niftis. There is an entry in the .nii header for the slope scale, but mri_convert appears to discard this value instead of copying it into the header (whereas mricron retains this value) We prefer to use mri_convert.
Is there a way to have mri_convert carry this value over? Why is it being discarded?
Thanks, Eli
(2005,100e) for MRI (Philips) (0028,1053) for PET (Philips)
I understand the PET may not be intended to work, but it doesn't seem to be copying over or applying the scale for either modality.
On 11/10/2016 12:40 PM, Douglas N Greve wrote:
What is the dicom tag for the scale?
On 11/10/2016 01:29 PM, Rockers, Elijah D. wrote:
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote:
> Hello, > > We have been using mri_convert to convert our dicoms into .nii > niftis. > There is an entry in the .nii header for the slope scale, but > mri_convert appears to discard this value instead of copying it > into the > header (whereas mricron retains this value) We prefer to use > mri_convert. > > Is there a way to have mri_convert carry this value over? Why is it > being discarded? > > Thanks, > Eli >
Doug,
Any update on this?
Thanks, Eli
On 11/10/2016 12:52 PM, Eli wrote:
(2005,100e) for MRI (Philips) (0028,1053) for PET (Philips)
I understand the PET may not be intended to work, but it doesn't seem to be copying over or applying the scale for either modality.
On 11/10/2016 12:40 PM, Douglas N Greve wrote:
What is the dicom tag for the scale?
On 11/10/2016 01:29 PM, Rockers, Elijah D. wrote:
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu Content-Type: text/plain; charset=utf-8; format=flowed I think we actually apply the scaling, so including it in the output would be inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote: >> Hello, >> >> We have been using mri_convert to convert our dicoms into .nii >> niftis. >> There is an entry in the .nii header for the slope scale, but >> mri_convert appears to discard this value instead of copying it >> into the >> header (whereas mricron retains this value) We prefer to use >> mri_convert. >> >> Is there a way to have mri_convert carry this value over? Why >> is it >> being discarded? >> >> Thanks, >> Eli >>
Try converting the PET after setting this env var
setenv FS_SLICE_LIN_SCALE 1
This causes mri_convert to read 28,1052 (intercept) and 28,1053 (slope). the output value should be
dicomvalue*slope + intercept
Is 2005,100e accompanied by an intercept ?
On 11/18/2016 11:25 AM, Rockers, Elijah D. wrote:
Doug,
Any update on this?
Thanks, Eli
On 11/10/2016 12:52 PM, Eli wrote:
(2005,100e) for MRI (Philips) (0028,1053) for PET (Philips)
I understand the PET may not be intended to work, but it doesn't seem to be copying over or applying the scale for either modality.
On 11/10/2016 12:40 PM, Douglas N Greve wrote:
What is the dicom tag for the scale?
On 11/10/2016 01:29 PM, Rockers, Elijah D. wrote:
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote:
Doug,
Can you verify? The dicom and mri_covnert .nii are identical (integer values, as per dicom convention) when we inspect the matrix in MATLAB.
The img matrices for mri_convert and dcm2nii are also identical, the difference appears to be that dcm2nii retains the slope scale in the header, allowing the viewer (freeview, pmod, etc..) to apply it.
We have been trying to nail down this issue for the last couple days, any insight would be appreciated.
Eli
On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu wrote: > Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N > Greve > greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert > slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: > 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu > Content-Type: text/plain; charset=utf-8; format=flowed I think we > actually apply the scaling, so including it in the output would be > inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote: >>> Hello, >>> >>> We have been using mri_convert to convert our dicoms into .nii >>> niftis. >>> There is an entry in the .nii header for the slope scale, but >>> mri_convert appears to discard this value instead of copying it >>> into the >>> header (whereas mricron retains this value) We prefer to use >>> mri_convert. >>> >>> Is there a way to have mri_convert carry this value over? Why >>> is it >>> being discarded? >>> >>> Thanks, >>> Eli >>>
Yes, (2005,100d) is the philips MRI intercept tag. it is zero in all of our scans.... unfortunately they are using private tags for philips MR. the 28,1052/1053 are the public tags, which their PET thankfully uses... but for the MR they are empty.
I'm using bash, so i did export FS_SLICE_LIN_SCALE=1, but there was no change in the output, still only seeing raw DICOM values. (also, slope scale did not copy into the header, either behavior would be desireable)
One more thing, the slope scale for Philips MRI, 2005,100e is actually an inverse slope scale, i.e., for accurate quantitative analysis, the images must be divided by the value in the slope scale header.... we are under the impression this is an issue unique to Philips MRI images. See https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3998685/ for more info.
On 11/18/2016 10:37 AM, Douglas N Greve wrote:
Try converting the PET after setting this env var
setenv FS_SLICE_LIN_SCALE 1
This causes mri_convert to read 28,1052 (intercept) and 28,1053 (slope). the output value should be
dicomvalue*slope + intercept
Is 2005,100e accompanied by an intercept ?
On 11/18/2016 11:25 AM, Rockers, Elijah D. wrote:
Doug,
Any update on this?
Thanks, Eli
On 11/10/2016 12:52 PM, Eli wrote:
(2005,100e) for MRI (Philips) (0028,1053) for PET (Philips)
I understand the PET may not be intended to work, but it doesn't seem to be copying over or applying the scale for either modality.
On 11/10/2016 12:40 PM, Douglas N Greve wrote:
What is the dicom tag for the scale?
On 11/10/2016 01:29 PM, Rockers, Elijah D. wrote:
As far as I can tell, mri_convert doesn't apply the scaling factor, and also doesn't copy over the slope scale header entry, when converting from dicom to nifti.
By contrast, dcm2nii does not apply the scale either, however, it copies over the slope scale header entry, enabling other software (eg Freeview..) to apply the scaling..
Hope this clarifies. Thanks for checking into this.
On 11/10/2016 12:18 PM, Douglas N Greve wrote:
Sorry, is it a scale in the dicom file or the nifti file that is not being applied properly?
On 11/10/2016 11:01 AM, Rockers, Elijah D. wrote: > Doug, > > Can you verify? The dicom and mri_covnert .nii are identical > (integer > values, as per dicom convention) when we inspect the matrix in > MATLAB. > > The img matrices for mri_convert and dcm2nii are also identical, > the > difference appears to be that dcm2nii retains the slope scale in > the > header, allowing the viewer (freeview, pmod, etc..) to apply it. > > We have been trying to nail down this issue for the last couple > days, > any insight would be appreciated. > > Eli > > On 11/09/2016 05:51 PM, freesurfer-request@nmr.mgh.harvard.edu > wrote: >> Message: 2 Date: Wed, 9 Nov 2016 16:01:18 -0500 From: Douglas N >> Greve >> greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] mri_convert >> slope scale To: freesurfer@nmr.mgh.harvard.edu Message-ID: >> 908818ac-156f-f6d2-1338-694626cde58f@nmr.mgh.harvard.edu >> Content-Type: text/plain; charset=utf-8; format=flowed I think we >> actually apply the scaling, so including it in the output would be >> inappropriate. On 11/09/2016 03:03 PM, Rockers, Elijah D. wrote: >>>> Hello, >>>> >>>> We have been using mri_convert to convert our dicoms into .nii >>>> niftis. >>>> There is an entry in the .nii header for the slope scale, but >>>> mri_convert appears to discard this value instead of copying it >>>> into the >>>> header (whereas mricron retains this value) We prefer to use >>>> mri_convert. >>>> >>>> Is there a way to have mri_convert carry this value over? Why >>>> is it >>>> being discarded? >>>> >>>> Thanks, >>>> Eli >>>>
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Rockers, Elijah D.