[MINC-users] SNR question

[Andrew Janke]

> Is there a minc tool to estimate the SNR of an anatomical volume
> automatically?  I have a T2 volume of the human brain and would like
> to have an approximation of the SNR.

Nothing automatic that I know of, but....

> If no tool exist can I do the following:
> -take a ROI in the background and compute the standard deviation of
> the signal in that ROI, std_background
> -take a ROI in the brain and find the std_data
> -SNR = std_data / std_background
>
> The obvious question is how big and in which area should my data ROI
> be.  Containing both white matter and gray matter? or it doesn't
> really matter to obtain a rough estimation of the SNR.

The above approach will work.  As for the size of the ROI, you should
find than an ROI around 5000 voxels should do the trick.

--
Andrew Janke      (a.janke at gmail.com || www.cmr.uq.edu.au/~rotor)
Canada->Montreal                                   Cell: +1 (514) 924 2012