search

DrSAR / SARlabpy

git clone git@pfeifer.phas.ubc.ca:SARlabpy (do not push to github, please)
http://code.SARlab.ca
Other
1 stars 0 forks source link

fast SNR estimate #292

Closed DrSAR closed 9 years ago

DrSAR commented 9 years ago

If a scan with a time-dimension exists we can calculate the SNR based on a ratio of mean to standard deviation for every pixel in the image. This has the advantage over ROI-based SNR's of incorporating artefacts in the areas where we falsely suspect signal voids.

firasm commented 9 years ago

Images from preliminary analysis of SNR - we actually do pretty well with SNR if we put the mouse tumour on top of the coil:

A. HPGS4 (lowest SNR yet) a-hpgs4h02 6r1 28

_B. HerS10 _ b-hers10do05 l01 7

C. NecS3 c-necs3hs05 ik1 11

_D. HerS11 _ d-hers11bs10 md1 6

E. HPGS13 - Test mouse Note that the two tumours are way far from the area where the receive coil is most sensitive - we really shot ourselves in the foot here by our positioning. Also, will you look at THAT - there is an SNR improvement of a factor of 10 that we can get from this, just by positioning!

e-hpgs13testmouse sv1 8

firasm commented 9 years ago

okay maybe not factor of 10, maybe factor of 5