Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

David Qixiang Chen
Hi all,
I've been trying to use the DWI denoising filters to filter my aquisitions.
I've found that the results always have large amount of black noise on
them, rendering the results unusable.
Just wondering if this is a known issue, is there anyway around this?
Attached is the screen shot.

My aquisition parameters:
GE 3T Signa HDx scanner, 128x128 matrix interpolated to 256x256.
60 directions. 47 slices. Image spacing: 0.9375x0.9374. Slice thickness: 2.7.
b=3000 and NEX=1

On a related issue. I want to resample the DWI volume to isovoxel
dimensions, but the DWI resample filter always fails in 3.6.
What is the recommended steps for doing this ?

Thanks,

- David Qixiang Chen

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject

Screenshot.png (110K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

Demian Wassermann-2
Hi David,

How are you doing?

Is there a way I can get the nrrd file of your DWI image? I would like to analyze it to help you

Thanks
Demian
--
Demian Wassermann, PhD
[hidden email]
LMI / PNL / SPL Labs
Harvard Medical School
Brigham and Women's Hospital
1249 Boylston, Boston, MA, USA

On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:

> <Attached Message Part.txt>

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

David Qixiang Chen
Hi Damian,

Here's the eddy corrected scan using tend epireg.
http://drop.io/eyqpfyi
http://drop.io/y0ydghc

Btw any ideas about doing isovoxel resampling on DWI images?

Thanks,

- David

On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
<[hidden email]> wrote:

> Hi David,
>
> How are you doing?
>
> Is there a way I can get the nrrd file of your DWI image? I would like to analyze it to help you
>
> Thanks
> Demian
> --
> Demian Wassermann, PhD
> [hidden email]
> LMI / PNL / SPL Labs
> Harvard Medical School
> Brigham and Women's Hospital
> 1249 Boylston, Boston, MA, USA
>
> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>
>> <Attached Message Part.txt>
>
>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

Demian Wassermann-2
Hi David,

I've been looking at your data trying to reproduce the problem and I wasn't able to.

Could you give me more details? Slicer version, filter that you are using, etc etc


Thanks
Demian


--
Demian Wassermann, PhD
[hidden email]
LMI / PNL / SPL Labs
Harvard Medical School
Brigham and Women's Hospital
1249 Boylston, Boston, MA, USA

On Jul 19, 2010, at 11:24 AM, David Qixiang Chen wrote:

> Hi Damian,
>
> Here's the eddy corrected scan using tend epireg.
> http://drop.io/eyqpfyi
> http://drop.io/y0ydghc
>
> Btw any ideas about doing isovoxel resampling on DWI images?
>
> Thanks,
>
> - David
>
> On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
> <[hidden email]> wrote:
>> Hi David,
>>
>> How are you doing?
>>
>> Is there a way I can get the nrrd file of your DWI image? I would like to analyze it to help you
>>
>> Thanks
>> Demian
>> --
>> Demian Wassermann, PhD
>> [hidden email]
>> LMI / PNL / SPL Labs
>> Harvard Medical School
>> Brigham and Women's Hospital
>> 1249 Boylston, Boston, MA, USA
>>
>> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>>
>>> <Attached Message Part.txt>
>>
>>

_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

David Qixiang Chen
Hi Demian,

This is consistently reproducible for me since the Joint Rician LMMSE
filter was introduced.
I've reproduced this in 3.7 alpha now.

I use the default settings:
estimation radius: 2,2,1
filtering radius: 2,2,1
# of neighboring gradients: 0

If you look at the individual gradients of the DWI volume, the black
specs are always there.
Can you tell me what settings you used? If you can't reproduced this,
that certainly is good news.

- David

On Tue, Jul 27, 2010 at 7:20 AM, Demian Wassermann
<[hidden email]> wrote:

> Hi David,
>
> I've been looking at your data trying to reproduce the problem and I wasn't able to.
>
> Could you give me more details? Slicer version, filter that you are using, etc etc
>
>
> Thanks
> Demian
>
>
> --
> Demian Wassermann, PhD
> [hidden email]
> LMI / PNL / SPL Labs
> Harvard Medical School
> Brigham and Women's Hospital
> 1249 Boylston, Boston, MA, USA
>
> On Jul 19, 2010, at 11:24 AM, David Qixiang Chen wrote:
>
>> Hi Damian,
>>
>> Here's the eddy corrected scan using tend epireg.
>> http://drop.io/eyqpfyi
>> http://drop.io/y0ydghc
>>
>> Btw any ideas about doing isovoxel resampling on DWI images?
>>
>> Thanks,
>>
>> - David
>>
>> On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
>> <[hidden email]> wrote:
>>> Hi David,
>>>
>>> How are you doing?
>>>
>>> Is there a way I can get the nrrd file of your DWI image? I would like to analyze it to help you
>>>
>>> Thanks
>>> Demian
>>> --
>>> Demian Wassermann, PhD
>>> [hidden email]
>>> LMI / PNL / SPL Labs
>>> Harvard Medical School
>>> Brigham and Women's Hospital
>>> 1249 Boylston, Boston, MA, USA
>>>
>>> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>>>
>>>> <Attached Message Part.txt>
>>>
>>>
>
>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

pieper
Administrator
Hi David -

Can you try with the dwi data from the tutorial?

http://www.slicer.org/slicerWiki/images/c/cf/DiffusionDataset.zip

I just tested this and didn't get any black spots.  I'm thinking the
issue may be overflow related to the dwi values in a particular dataset.

-Steve

On Aug/5/10 11:42 AM, David Qixiang Chen wrote:

> Hi Demian,
>
> This is consistently reproducible for me since the Joint Rician LMMSE
> filter was introduced.
> I've reproduced this in 3.7 alpha now.
>
> I use the default settings:
> estimation radius: 2,2,1
> filtering radius: 2,2,1
> # of neighboring gradients: 0
>
> If you look at the individual gradients of the DWI volume, the black
> specs are always there.
> Can you tell me what settings you used? If you can't reproduced this,
> that certainly is good news.
>
> - David
>
> On Tue, Jul 27, 2010 at 7:20 AM, Demian Wassermann
> <[hidden email]>  wrote:
>> Hi David,
>>
>> I've been looking at your data trying to reproduce the problem and I wasn't able to.
>>
>> Could you give me more details? Slicer version, filter that you are using, etc etc
>>
>>
>> Thanks
>> Demian
>>
>>
>> --
>> Demian Wassermann, PhD
>> [hidden email]
>> LMI / PNL / SPL Labs
>> Harvard Medical School
>> Brigham and Women's Hospital
>> 1249 Boylston, Boston, MA, USA
>>
>> On Jul 19, 2010, at 11:24 AM, David Qixiang Chen wrote:
>>
>>> Hi Damian,
>>>
>>> Here's the eddy corrected scan using tend epireg.
>>> http://drop.io/eyqpfyi
>>> http://drop.io/y0ydghc
>>>
>>> Btw any ideas about doing isovoxel resampling on DWI images?
>>>
>>> Thanks,
>>>
>>> - David
>>>
>>> On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
>>> <[hidden email]>  wrote:
>>>> Hi David,
>>>>
>>>> How are you doing?
>>>>
>>>> Is there a way I can get the nrrd file of your DWI image? I would like to analyze it to help you
>>>>
>>>> Thanks
>>>> Demian
>>>> --
>>>> Demian Wassermann, PhD
>>>> [hidden email]
>>>> LMI / PNL / SPL Labs
>>>> Harvard Medical School
>>>> Brigham and Women's Hospital
>>>> 1249 Boylston, Boston, MA, USA
>>>>
>>>> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>>>>
>>>>> <Attached Message Part.txt>
>>>>
>>>>
>>
>>
> _______________________________________________
> slicer-users mailing list
> [hidden email]
> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
> To unsubscribe: send email to [hidden email] with unsubscribe as the subject
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

David Qixiang Chen
Hi Steve,
I just tested this with the tutorial data, no black spots.
Have you tried it on my sample data? I'm on a 64bit system btw.

Looks like it's limited to my scans. Could this be due to the large
number of directions I have? Our scans are usually done with either 25
directions or 60 directions.

Thanks,

- David

On Thu, Aug 5, 2010 at 12:39 PM, Steve Pieper <[hidden email]> wrote:

> Hi David -
>
> Can you try with the dwi data from the tutorial?
>
> http://www.slicer.org/slicerWiki/images/c/cf/DiffusionDataset.zip
>
> I just tested this and didn't get any black spots.  I'm thinking the issue
> may be overflow related to the dwi values in a particular dataset.
>
> -Steve
>
> On Aug/5/10 11:42 AM, David Qixiang Chen wrote:
>>
>> Hi Demian,
>>
>> This is consistently reproducible for me since the Joint Rician LMMSE
>> filter was introduced.
>> I've reproduced this in 3.7 alpha now.
>>
>> I use the default settings:
>> estimation radius: 2,2,1
>> filtering radius: 2,2,1
>> # of neighboring gradients: 0
>>
>> If you look at the individual gradients of the DWI volume, the black
>> specs are always there.
>> Can you tell me what settings you used? If you can't reproduced this,
>> that certainly is good news.
>>
>> - David
>>
>> On Tue, Jul 27, 2010 at 7:20 AM, Demian Wassermann
>> <[hidden email]>  wrote:
>>>
>>> Hi David,
>>>
>>> I've been looking at your data trying to reproduce the problem and I
>>> wasn't able to.
>>>
>>> Could you give me more details? Slicer version, filter that you are
>>> using, etc etc
>>>
>>>
>>> Thanks
>>> Demian
>>>
>>>
>>> --
>>> Demian Wassermann, PhD
>>> [hidden email]
>>> LMI / PNL / SPL Labs
>>> Harvard Medical School
>>> Brigham and Women's Hospital
>>> 1249 Boylston, Boston, MA, USA
>>>
>>> On Jul 19, 2010, at 11:24 AM, David Qixiang Chen wrote:
>>>
>>>> Hi Damian,
>>>>
>>>> Here's the eddy corrected scan using tend epireg.
>>>> http://drop.io/eyqpfyi
>>>> http://drop.io/y0ydghc
>>>>
>>>> Btw any ideas about doing isovoxel resampling on DWI images?
>>>>
>>>> Thanks,
>>>>
>>>> - David
>>>>
>>>> On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
>>>> <[hidden email]>  wrote:
>>>>>
>>>>> Hi David,
>>>>>
>>>>> How are you doing?
>>>>>
>>>>> Is there a way I can get the nrrd file of your DWI image? I would like
>>>>> to analyze it to help you
>>>>>
>>>>> Thanks
>>>>> Demian
>>>>> --
>>>>> Demian Wassermann, PhD
>>>>> [hidden email]
>>>>> LMI / PNL / SPL Labs
>>>>> Harvard Medical School
>>>>> Brigham and Women's Hospital
>>>>> 1249 Boylston, Boston, MA, USA
>>>>>
>>>>> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>>>>>
>>>>>> <Attached Message Part.txt>
>>>>>
>>>>>
>>>
>>>
>> _______________________________________________
>> slicer-users mailing list
>> [hidden email]
>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>> To unsubscribe: send email to
>> [hidden email] with unsubscribe as the
>> subject
>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject
Reply | Threaded
Open this post in threaded view
|

Re: Black noise artifact on Joint Rician LMMSE filter and resample DWI to isovoxel

pieper
Administrator
Hi David -

I didn't try your data (looks like the link is no longer active).
Demian, did you try with David's data?

-Steve

On Aug/5/10 12:50 PM, David Qixiang Chen wrote:

> Hi Steve,
> I just tested this with the tutorial data, no black spots.
> Have you tried it on my sample data? I'm on a 64bit system btw.
>
> Looks like it's limited to my scans. Could this be due to the large
> number of directions I have? Our scans are usually done with either 25
> directions or 60 directions.
>
> Thanks,
>
> - David
>
> On Thu, Aug 5, 2010 at 12:39 PM, Steve Pieper<[hidden email]>  wrote:
>> Hi David -
>>
>> Can you try with the dwi data from the tutorial?
>>
>> http://www.slicer.org/slicerWiki/images/c/cf/DiffusionDataset.zip
>>
>> I just tested this and didn't get any black spots.  I'm thinking the issue
>> may be overflow related to the dwi values in a particular dataset.
>>
>> -Steve
>>
>> On Aug/5/10 11:42 AM, David Qixiang Chen wrote:
>>>
>>> Hi Demian,
>>>
>>> This is consistently reproducible for me since the Joint Rician LMMSE
>>> filter was introduced.
>>> I've reproduced this in 3.7 alpha now.
>>>
>>> I use the default settings:
>>> estimation radius: 2,2,1
>>> filtering radius: 2,2,1
>>> # of neighboring gradients: 0
>>>
>>> If you look at the individual gradients of the DWI volume, the black
>>> specs are always there.
>>> Can you tell me what settings you used? If you can't reproduced this,
>>> that certainly is good news.
>>>
>>> - David
>>>
>>> On Tue, Jul 27, 2010 at 7:20 AM, Demian Wassermann
>>> <[hidden email]>    wrote:
>>>>
>>>> Hi David,
>>>>
>>>> I've been looking at your data trying to reproduce the problem and I
>>>> wasn't able to.
>>>>
>>>> Could you give me more details? Slicer version, filter that you are
>>>> using, etc etc
>>>>
>>>>
>>>> Thanks
>>>> Demian
>>>>
>>>>
>>>> --
>>>> Demian Wassermann, PhD
>>>> [hidden email]
>>>> LMI / PNL / SPL Labs
>>>> Harvard Medical School
>>>> Brigham and Women's Hospital
>>>> 1249 Boylston, Boston, MA, USA
>>>>
>>>> On Jul 19, 2010, at 11:24 AM, David Qixiang Chen wrote:
>>>>
>>>>> Hi Damian,
>>>>>
>>>>> Here's the eddy corrected scan using tend epireg.
>>>>> http://drop.io/eyqpfyi
>>>>> http://drop.io/y0ydghc
>>>>>
>>>>> Btw any ideas about doing isovoxel resampling on DWI images?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> - David
>>>>>
>>>>> On Sat, Jul 17, 2010 at 2:08 PM, Demian Wassermann
>>>>> <[hidden email]>    wrote:
>>>>>>
>>>>>> Hi David,
>>>>>>
>>>>>> How are you doing?
>>>>>>
>>>>>> Is there a way I can get the nrrd file of your DWI image? I would like
>>>>>> to analyze it to help you
>>>>>>
>>>>>> Thanks
>>>>>> Demian
>>>>>> --
>>>>>> Demian Wassermann, PhD
>>>>>> [hidden email]
>>>>>> LMI / PNL / SPL Labs
>>>>>> Harvard Medical School
>>>>>> Brigham and Women's Hospital
>>>>>> 1249 Boylston, Boston, MA, USA
>>>>>>
>>>>>> On Jul 17, 2010, at 11:54 AM, Ron Kikinis wrote:
>>>>>>
>>>>>>> <Attached Message Part.txt>
>>>>>>
>>>>>>
>>>>
>>>>
>>> _______________________________________________
>>> slicer-users mailing list
>>> [hidden email]
>>> http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
>>> To unsubscribe: send email to
>>> [hidden email] with unsubscribe as the
>>> subject
>>
_______________________________________________
slicer-users mailing list
[hidden email]
http://massmail.spl.harvard.edu/mailman/listinfo/slicer-users
To unsubscribe: send email to [hidden email] with unsubscribe as the subject